liquidrocketmetaltanks andtankcomponentstank configuration vehicle-tank optimization subsyste m-tank...
TRANSCRIPT
NASA NASA SP-8088
SPACEVEHICLE
DESIGNCRITERIA
(CHEMICALPROPULSION)
LIQUIDROCKETMETALTANKS
ANDTANKCOMPONENTS
MAY1974
NATIONAL AERONAUTICS AND SPACE ADMINISTRATION
https://ntrs.nasa.gov/search.jsp?R=19750004950 2020-03-22T09:21:27+00:00Z
FOREWORD
NASA experience has indicated a need for uniforn3 criteria for the design of space vehicles.
Accordingly, criteria are being developed ill the following areas of technology :
Environment
Structures
Guidance and ('ontrol
Chemical Propulsion
Individual components of this work will bc isstied as separate monographs as soon as they
are completed. This document, part of the series on Chemical Propulsion, is one such
monograph. A list of all monographs isstied to date can be lkmnd on the final pages of this
docu lllCll t.
These monographs are to be regarded as guides to design and not as NASA requirements,
except as may be specified in formal project specifications. It is expected, however, that
these docunlents, revised as experience may indicate to be desirable, eventually will provide
uniform design practices for NASA space vehicles.
This monograph, "Liquid Rocket Metal Tanks and Tank ('omponents", was prepared under
the direction of Howard W. l)ouglass, Chief, l)esign Criteria Office, Lewis Research Center:
project management was by M. Murray Bailey. The monograph was written by W. A. Wagner
of Space Division, Rockwell International Corporation, and was edited by Russell B. Keller,
Jr. of Lewis. Significant contributions to the text were made by C. I). Brownfield, Space
Division, Rockwell International Corporation. To assure teclmical accuracy of this
document, scientists and engineers throughout the technical community participated in
interviews, consultations, and critical review of the text. In particular, Richard A.
Morehouse of The Boeing Company: Fred R. Schwartzberg of Martin Marietta Company;
Leo M. Thompson of Bell Aerospace Company, Division of Textron; and Gordon T. Smithand Richard T. Barrett of the Lewis Research Center individually and collectively reviewed
the monograph in detail.
Comments concerning the technical content of this monograph will be welcomed by the
National Aeronautics and Space Administration, Lewis Research Center (Design Criteria
Office), Cleveland, Ohio 44135.
May 1974
GUIDE TO THE USE OF THIS MONOGRAPH
The purpose of this monograph is to organize and present, for effective use in design, tile
significant experience and knowledge accmnulated in development and operational
progranls to date. It reviews and assesses current design practices, and from them establishes
firm guidance for achieving greater consistency in design, increased reliability in the end
product, and greater efficiency in the design effort. The monograph is organized into two
major sections that are preceded by a brief introduction and complemented by a set of
rcferences.
The State of the Art, section 2, reviews and discusses the total design problem, and
identifies which design elements are involved in successful design. It describes succinctly the
current tecnnology pertaining to these elements. When detailed information is required, the
best available references are cited. This section serves as a survey of the subject that provides
background material and prepares a proper technological base for the Design Criteria and
Recommended Practices.
The Desig;z Criteria, shown in italics in section 3, state clearly and briefly what rule, guide,
limitation, or standard must be imposed on each essential design element to assure
successful design. The Design Criteria can serve effectively as a checklist of rules for the
project manager to use in guiding a design or in assessing its adequacy.
The Recommended Practices, also in section 3, state how to satisfy each of the criteria.
Whenever possible, the best procedure is described; when this cannot be done concisely,
appropriate references are provided. The Recommended Practices, in conjunction with the
Desig;t Criteria. provide positive guidance to the practicing designer on how to achieve
successful design.
Both sections have been organized into decimally numbered subsections so that the subjects
within similarly numbered subsections correspond from section to section. The format for
the Contents displays this continuity of subject in such a way that a particular aspect of
design can be followed through both sections as a discrete subject.
The design criteria monograph is not intended to be a design handbook, a set of
specifications, or a design manual. It is a summary and a systematic ordering of the large and
loosely organized body of existing successful design techniques and practices. Its value and
its merit should be judged on how effectively it makes that material available to and useful
to the designer.
111
CONTENTS
I. iNTRODUCTION ............
2. STATE OF TIlE ART ...........
3. DESIGN CRITERIA and Recomn3ended Practices
APPENDIX A - Conversioh of U.S. Customary Units to Sl Units ............
APPENDIX B - Glossary ............................
REFERENCES ................................
NASA Space Vehicle Design Criteria Monographs Issued to Date .............
SU BJ ECT STATE OF THE ART DESIGN CRITERIA
TANK CONFIGURATION
Vehicle-Tank Optimization
Subsyste m-Tank Optimization
TANK MATERIAL
Mechanical Properties
Temperature Effects on Properties
Fatigue Strength
Creep
Biaxial-Stress Properties
Fabrication Considerations
Shaping and Forming
Welding
Thermal Processing
Material Compatibility with Environments
Sources of Material/Fluid Reactions
2.1 13 3.1 93
ZI.I 13 3.1.1 93
2.1.2 14 3.1.2 93
2.2 16 3.2 94
2.2.1 17 3.2.1 94
2.2.1.1 18 3.2.1.1 94
2.2.1.2 19 3.2.1.2 95
2.2.1.3 19 3.2.1.3 95
2,2.1.4 20 3.2.1.4 95
2.2.2 21 3.2.2 96
2.2.2.1 21 3.2.2.1 96
2.2.2.2 23 3.2.2.2 97
2.2.2.3 24 3.2.2.3 97
2.2.3 28 3.2.3 98
2.2.3. 1 30 3.2.3. I 99
Page
1
3
93
127
129
139
151
SUBJECT STATE OF THE ART DESIGN CRITERIA
Manufacturing Fluids and ProcessTesting FluidsStored Fluids
Atmospheric and EnvironmentalCorrosion
Types of Material/Fluid Reactions
Stress-Corrosion CrackingGalvanic Corrosion
Hydrogen-EnvironmentEmbrittlement
Material Ignition
Fracture Control
Material Fracture ToughnessSafe-Life Analysis
Additional Elements in a Fracture-
C ontrol Program
TANK STRUCTURAL DESIGN
Safety Factor
Loads Analysis
Tank Sidewall
End Closure
Intertank BulkheadAttachment
Membrane ThicknessSidewall
Skin-Stringer-FrameWaffle
Monocoque
End Closure
Forward Bulkhead
Aft Bulkhead
Intertank Bulkhead
Sandwich Construction
2.2.3.1.1 30
2.2.3.1.2 31
2.2.3.1.3 31
2.2.3.1.4 32
2.2.3.2 33
2.2.3.2.1 33
2.2.3.2.2 37
2.2.3.2.3 38
2.2.3.2.4 39
2.2.4 40
2.2.4.1 43
2.2.4.2 47
2.2.4.3 48
2.3 49
2.3.1 50
2.3.2 52
2.3.2.1 52
2.3.2.2 52
2.3.2.3 52
2.3.2.4 53
2.3.3 53
2.3.4 54
2.3.4.1 54
2.3.4.2 55
2.3.4.3 58
2.3.5 58
2.3.5.1 61
2.3.5.2 62
9_.3.5.3 64
3.2.3.1.1 99
3.2.3.1.2 1003.2.3.1.3 100
3.2.3.1.4 100
3.2.3.2 101
3.2.3.2.1 101
3.2.3.2.2 102
3.2.3.2.3 102
3.2.3.2.4 102
3.2.4 103
3.2.4.1 104
3.2.4.2 105
3.2.4.3 105
3.3 106
3.3.1 106
3.3.2 107
3.3.2.1 1073.3.2.2 107
3.3.2.3 108
3.3.2.4 108
3.3.3 108
3.3.4 109
3.3.4.1 109
3.3.4.2 109
3.3.4.3 111
3.3.5 111
3.3.5.1 111
3.3.5.2 112
3.3.5.3 112
3.3.5.3.1 113
vi
SUBJECT STATEOFTHE ART DESIGN CRITERIA
Attachment Junctures
Weld Joints
Bulkhead/Sidewall Juncture
Bosses and Support Provisions
Openings and Access Doors
TANK COMPONENTS
Propellant Slosh and Vortex SuppressionDevices
Propellant Positioning Devices
Propellant Expulsion Devices
Expulsion EfficiencyMaterial
Design Margin
Tank Insulation
TANK DESIGN ANALYSIS
Strength Analysis
Tension-Loaded Structure
Compression-Loaded Structure
Major JuncturesLocal Attachments and Openings
Structural Dynamics
Bending Frequency
External Dynamic Environment
TANK FABRICATION
TESTING AND INSPECTION
2.3.6 65 3.3.6 113
2.3.6.1 65 3.3.6.1 1132.3.6.2 67 3.3.6.2 115
2.3.6.3 69 3.3.6.3 115
2.3.7 71 3.3.7 116
2.4 72 3.4 118
2.4.1 72
2.4.2 72
2.4.3 74
2.4.3.1 76
2.4.3.2 77
2.4.3.3 79
2.4.4 82
2.5 82
2.5.1 85
2.5.1.1 86
2.5.1.2 87
2.5.1.3 87
2.5.1.4 88
2.5.2 88
2.5.2.1 89
2.5.2.2 89
2.6 89
2. 7 90
3.4.1 118
3.4.2 118
3.4.3 119
3.4.3.1 119
3.4.3.2 120
3.4.3.3 120
3.4.4 121
3.5 121
3.5.1 121
3.5.1.1 122
3.5.1.2 122
3.5.1.3 122
3.5.1.4 123
3.5.2 123
3.5.2.1 124
3.5.2.2 124
3.6 125
3. 7 125
vii
Figure
l
2
3
LIST OF FIGURES
Title Page
Exploded view of major components for Saturn IC booster ............ 10
Exploded view of major components for Saturn S-II stage ............. I 1
Sketch illustrating surface-crack model and
nomenclature used in fracture-mechanics analysis ............... 41
4 Material fracture toughness vs yield strength,
density' normalized .......................... 44
5 Two kinds of skin-stringer-frame construction .................. 55
6 Sidewall construction on LH 2 tank for Saturn S-II stage .............. 56
7 Two basic waffle configurations ....................... 57
8 Some surfaces of revolution that have been
analyzed by computer subroutines ...................... 59
9 Cutaway drawing of the LOX/LH 2 tanks in the Saturn
S-II stage ................................ 60
10 Example of honeycomb sandwich structure .................. 61
I l Variation of tank weight with variation in end-closure configuration
(LH 2 tanks for S-II stage) ......................... 63
12 Sketches of two basic types of intertank bulkheads ................ 64
13 Weld joint configurations for vehicle tanks ................... 65
14 Weld joint configuralions for subsystem tanks .................. 66
15 Examples of Y-ring bulkhead/sidewall junctures ................. 68
16 Example of bulkhead/sidewall juncture used in Centaur .............. 68
17 Three designs for a tank access door ..................... 71
18 Schematics of two basic types of surface-tension
devices for fluid control .......................... 73
viii
Figure
19
20
21
22
23
24
25
26
27
Title Page
Schematics of three kinds of positive expulsion devices .............. 75
Examples of internal and external insulation .................. 83
Jettisonable insulation panels (Centaur) .................... 84
Insulation of intertank bulkhead (Centaur) ................... 85
Realignment of an eccentric weld joint under load ................ 86
Structural elements of bulkhead/sidewall junctures
subjected to analysis ........................... 88
Sketch of geometry of a weld joint for a subsystem tank ............. 114
Two configurations of structure for attachments to a tank ............. 116
Recommended geometry for transition from boss totank membrane .............................. 117
ix
Table
1
II
Ill
IV
V
VI
VI1
VIII
LIST OF TABLES
Title Page
Chief Design Features of Vehicle Tanks ................... 4
5Chief Design Features of Subsystem Tanks .................
8Chief Design Features of Positive-Expulsion Tanks ..............
Comparison of Fabrication Methods for Pressure-Vessel Components ....... 22
Thermal Processes Applicable to the Construction of Metal Tanks ........ 26
Causes and Effects of Metal/Fluid Reactions ................. 29
Comparison of the Resistance of Various Alloys to Stress-Corrosion35
Cracking in Natural Environments .....................
Sample Trade Study for Selection of Optimum Sidewall Configuration ...... 110
x
LIQUID ROCKET METAL TANKS
AND TANK COMPONENTS
1. INTRODUCTION
The flightweight tanks containing various system fluids are an important part of any liquid
propellant rocket propulsion system. They vary in size and shape from a cylindrical tankholding many thousands of gallons of booster propellant to a sphere holding only a few
cubic feet of spacecraft pressurant gas at high pressure. The continuous improvement and
upgrading of the various codes governing pressure-vessel design and construction are
indicative of the emphasis and concern placed on tank design because of the explosive
hazard of tankage even under moderate pressures. However, additional guidelines and
practices are required to ensure that tanks for aerospace applications are of optimum design.
This monograph has been prepared to delineate the significant guidelines and practices for
successful design of aerospace tanks and tank components such as expulsion devices,
standpipes, and baffles.
The structural-weight efficiency of aerospace tanks strongly influences the payload
capabilities. The incentive to minimize tank weight by use of high-strength, brittle materials
operated at a high fraction of yield strength must be balanced against the reliability
requirements and economic constraints that are inherent in each particular design situation.Many metal alloys under high stress are sensitive both to small inherent flaws and to the
effects of various external environments. Flaw growth induced by stress or by
environmental conditions has led to tank rupture even at normal operational pressure.
Fracture-control methods based on the recently developed technology of linear-elastic
fracture mechanics provide a means for minimizing such failures.
Failures of tank assembly components, although usually not as ominous as a tank rupture,
have just as surely led to mission failures; for example, expulsion devices and standpipes
have failed, thereby preventing proper propellant consumption. The designer therefore must
employ the same care for components as for tanks in establishing strength margins, selectingmaterial, and allowing for environmental effects.
The material in this monograph is organized around the major considerations in the design
of metal tanks. Although these considerations are listed as separate entities, they are
interrelatedto varying degrees;and thesevariousinteractionsarediscussed.Becauseofextremedifferencesin structural complexity,vehicle tanks are treatedseparatelyfromsubsystemtanks.Vehicle tanks are tanks that serve both as primary integral structure of a
vehicle and as a container of pressurized propellants. Subsystem tanks are containers ofpressurized fluids or gases that are mounted internally in a vehicle, are essentially isolated
from adverse vehicle loads, and are of monocoque design.
In the development of a tank, the initial design activity is simply the determination of tank
shape or configuration within the constraints of mating vehicle structure or available
mounting space. When the basic configuration has been defined, the next activity is material
selection. Mechanical properties, fracture toughness, environmental compatibility, cost,availability, and fabrication factors must be considered in material selection. Detail tank and
component design follow the material selection. The objective in detail design is to satisfy
the tank volume and shape requirements with the selected material in an optimum manner.
The significant elements in detail tank design are vehicle-tank sidewall structure, weld joints
at bulkhead and attachment junctures, and ports and access openings. Additional design
considerations are the influence and effect of fabrication processes on tank and componentdesign and, finally, the testing and inspection that are required to establish confidence in a
tank design.
2. STATE OF THE ART
Over the past ten years, hundreds of lightweight, high-strength tanks have been developed
for use on liquid rocket propulsion systems. Reference 1 identifies over 75 different
programs in which tanks were used. Tables I, II, and III present some of the significantdesign characteristics of a representative cross section of tank designs dating back to the late
1950's. Such a multitude of programs obviously presents wide variations in factors such as
mission duration, mission environment, fluids employed, and fluid energy levels, each of
which influence the design of the tanks utilized in a given vehicle.
The largest tanks are the main tanks of the launch vehicles, which must withstand significant
compressive loads and flight-induced flexure as well as internal pressure and fluid slosh
loads. Typically, there is a single oxidizer tank and a single fuel tank in each main stage of
the overall vehicle. Figures 1 and 2 show an exploded view of the primary constituents of the
Saturn S-IC booster and S-II stage, respectively.
Each stage or spacecraft of a vehicle complex employs smaller tanks in various subsystems
such as reaction control, pressurization, and hydraulic. These tanks are internally mounted,
are usually of monocoque design, have integral fluid ports and support provisions, and
contain an expulsion device if liquid must be supplied under random low-g conditions.
The Atlas, which was the first booster of major size developed, is unique in that the
lightweight monocoque design of the vehicle tanks requires internal pressure to precludemembrane* buckling. The internal common bulkhead, which separates the forward LOX
tank and aft RP-1 tank, also requires a positive pressure on the fuel side to prevent
structural failure. Stringent weight limitations led to the use of 301 CRES in the extra full
hard (XFH) condition for these tanks. The successive cylindrical sections are overlapped,
joining being accomplished by spot welds; a seam weld is added at each lap joint to preventleakage. The longitudinal welds in a particular cylindrical segment are butt welds. The
majority of later booster designs are of waffle or frame/stringer sidewall design with tank
sections joined by butt welding. The membranes are machined from thick sheet or plate
stock, a process that enables incorporation of thickened weld lands, thickened accessory or
structural attach points, and stringers that are integral with the membrane.
For the pressure and load range on large-diameter boosters such as Saturn IC and Titan 1II,
hoop tensile forces dictate the membrane (skin) thickness, and additional material
(structure) is added as required for longitudinal compressive loads. Aluminum alloys
(predominantly 2014-T6 and 2219-T87) have been the primary choice of material because
they possess good strength-to-density ratios and excellent ductility and toughness at both
room and cryogenic temperatures. Although welding has presented problems, an increased
*Terms, symbols, materials, and abbreviations are defined or identified in Appendix B.
!i I tti
I
c_
ao
• d c_ " c; c_"
- _=_ _ __
_._ o_ o__ _
s©
o
x x _ x x ,_
o _ o _ _ As
_ _ _ _ _' _.
x x x
4
_ _i_i_i_i_ _i_i_i_i
- - o_ -
_ _ _
_ _ _ _ . _
_× -× -_ -
e-
E
o. _- _- _= _0 _ _ _
°
o
_o
O =
8
I
o
rn
q
_o_
c_ d
_ = = : o oo ..o
._ r--
E
oc5
S
E E
r--
z_ _ _ _ _ _
(.J _.j
_ 3 o
o.
I i 1 o o I I I i I !
, _ _ o o _ _ _
• . ,
>
.__ _ _ _ _ _ "_
E
Forward skirt
LOX tank
Intertankstructure
LO X feedlines (5)
Anti-vortexbaffles
Frames/baffles
Fueltank
Thrust structure
Figure 1. - Exploded view of major components for Saturn IC booster.
]0
/LH 2 tank
Forward skirt
LOX tank
Aft skirt and
thrust structure
Aft interstage
Figure 2. - Exploded view of major components for Saturn S-II
stage.
knowledge of proper tooling, environmental control, joint preparation, and welding
procedures as well as improved techniques for inspection, proof testing, and repair have
made it possible to produce consistently reliable welded joints in these materials.
Although vehicle tanks vary significantly in size, load per inch, and structural complexity,
all these tanks have complied with the following principles:
(1) Gross stress levels at proof and operating conditions have been maintained below
yield strength of the material.
11
(2) Local yielding is permittedarounddiscontinuitiesif structuralintegrity is notcompromised.
(3) The factorof safetyprovidesadequatemarginfor uncertaintiesinstressanalysis,applied loading, and fabricationand permitssuitablemarginfor unavoidablestrengthdegradationduringservicelife.
(4) The factor of safetyis basedprimarily onexperience,qualitativeassessmentofuncertaintiesof thespecificdesign,andreliability requirements.
(5) Fracturestrengthis greaterthanyield andequalto or greaterthan minimumguaranteedultimatestrength.
(6) Flawsor defectsarefoundby inspectionandrepairedif permissible.
For the subsystemtankssummarizedin table II, it is notablethat in none of the designfeatures shown are there data that indicate a changing trend in tank design. These tanks also
comply with the principles set forth above for vehicle tanks. The weld joints are exclusively
butt welds. The preweld configurations of the weld joints are predominantly "V" or "J"
groove for the thicker membrane tanks and "burndown" lips for the thinner membrane
tanks. With few exceptions, welding is accomplished by the tungsten-inert-gas (TIG)method.
Various types of positive expulsion devices have been used successfully (table Ill); each typeof device has advantages and disadvantages. Since the expulsion device must be compatible
with the ultimate tank usage requirements, no particular expulsion device can be considered
universally superior; thus, the degree of usage of a particular type of expulsion device does
not necessarily mean that it is the superior method. In the past decade, the advances in the
technology of positive-expulsion devices have been consistent with the increasing severity of
space mission environments. Problems of material compatibility with propellants,propellant/gas permeation through thin bladder membranes, operation at cryogenic
temperatures, and multicycle requirements, to name a few, have been solved. It can be
expected that this technology will continue to advance as more prolonged space
explorations are undertaken in the future.
In summary, the basic design approach for vehicle tanks, subsystem tanks, and expulsion
devices to a large degree has not changed significantly over the past decade. Considerable
development has occurred in the technologies associated with the production of lightweight,
high-quality pressure vessels, especially in large sizes, and in the techniques for ensuring that
these tanks meet the unusually high levels of structural reliability demanded in complex
aerospace vehicle systems. Advances have been most notable in the following areas:
(1) Development of fracture-mechanics concepts and of methods for applying these
concepts to pressure-vessel fracture control (refs. 2, 3, and 4)
12
(2) Developmentof alloyswith improvedstrength,toughness,andfabricabitity
(3) Developmentof informationon interactioneffectsbetweenstructuralmaterialsandenvironments
(4) Developmentof equipmentand fabrication techniquescapableof producing,forming,andweldinglarge,thin-wallcomponents
(5) Developmentof improved NDI (nondestructiveinspection) techniquesandequipment.
Althoughadvancesin fracturemechanicsandNDI techniqueshavebeenrealized,theexactnatureof applicationto suchbasictank designdecisionsasmaterialselection,stresslevelselection,anddefinitionof proof-testrequirementsremainsasubjectof currentcontroversyandconsiderablemisunderstandingamongstructuraldesigners.
2.1 TANK CONFIGURATION
In the first phase of design, where the shape and size of the tank are established, vehicle
tanks and subsystem tanks present significantly different problems to the designer. Vehicle
tanks are an integral part of the vehicle structure and must sustain the compressive loads of
the overall vehicle stack. They are operated at comparatively low internal pressures (usually
less than 100 psig) and must be dimensionally compatible with the adjacent stage or
payload. Extensive consideration must be given to bulkhead shape because it affects tank
length and the requirements for structural stiffening, the extent depending on material used.
Subsystem tanks usually are of monocoque design, are internally mounted within the
vehicle, usually operate at high stress levels, and are insulated from vehicle structuredeflection by appropriately designed mountings. Except where problems in installation
space arise, the tank shape usually is determined on the basis of structural efficiency.
2.1.1 Vehicle-Tank Optimization
The design of a vehicle tank begins with consideration of the entire launch vehicle in terms
of mission requirements. The initial study often is accomplished with the aid of
computerized vehicle-synthesis programs (ref. 5), which perform extensive preliminarystudies designed to establish overall vehicle configuration and to provide constraints for the
detail design. A large number of options are evaluated to arrive at the final vehicle
configuration. These options include the number of stages; the number, size, and type of
engine for each stage; the selection and location of propellants for each stage; and the
13
length, diameter, and materialof each tank. Preliminarydesignstudiesare iterativeprocesses,and the designor structuralanalysisis limited to what is necessaryto establishfeasibilityand to arriveat a reasonablyaccuratedeterminationof masspropertiessuchascenterof gravity,centerof pressure,massdistribution,momentof inertia,andtotal weight.Thesepropertiesare integratedwith aerodynamicforcesand tank internalpressurestodetermine vehicle net shear, bending, and axial loads. The vehicle structural configuration is
modified to meet these requirements and the process iterated until it cenverges to anoptimum vehicle design.
The final output of the preliminary design provides the constraints for the next step: the
detail design and analysis. These constraints are definition of the tank diameter and length,tank internal pressure requirements, and propellant type, location, and volume. Additional
constraints are added by program management. These constraints, based on the type ofmission coupled with judgment and past experience, are the factor of safety; criteria for
failure mode of tank pressure regulator or vent valve; method of combining loads; and "free
standing" capability on the launch pad (unpressurized) with any or all other stages full orempty.
2.1.2 Subsystem-Tank Optimization
As in the case of the vehicle tanks, the primary design objective for the smaller subsystem
tanks usually is optimum design in terms of minimum weight or minimum design marginwithout impairment of reliability. For the smaller subsystem tanks, however, it should be
emphasized that an overview of costs for a liquid rocket vehicle program (with multiple
subsystems) may impose considerations of using existing forging dies or may require tankdesign commonality between subsystems; either requirement preempts the minimum
weight/margin goal. For purposes of this monograph, however, only the optimization ofdesign for a new, minimum-weight tank will be discussed.
Unlike vehicle tanks, which invariably are large liquid-carrying tanks, subsystem tanks may
be designed in any one of three configurations: liquid-carrying only, liquid-carrying with a
positive expulsion device, and gas-carrying. Each type presents distinct design problems.
Liquid-carrying tanks. - Optimization of a liquid-carrying tank is relatively easy because thefluids are considered incompressible, and therefore usable fluid volume and hence tank
volume becomes a constant. It is advantageous to keep the liquid working pressure at thelowest possible value that permits minimum wall thicknesses. Once the basic fluid volume to
accomplish a vehicle mission is established, the various delta volumes typical of most
operating rocket systems are identified and added. Since fluids expand as temperatureincreases, ullage volumes consistent with the predicted usage environment are added.
Loading errors, fluid displacement of any internal structure and accessories, and fluid traps
14
arecompensatedfor by increasedtank size.If the rocketsystemisbipropellant,feedoutimbalanceis considered;and,if acryogenicfluid is involved,theattendantboiloff loss must
be offset by an appropriate increment during filling.
Following the identification of all volume increments and the determination of total fluidvolume, selection of tank shape is the next decision point. Cylinders with llemispherical
ends and spheres are the most common shapes for the smaller subsystem tanks. Limitations
of installation space and mounting difficulties frequently eliminate the spherical tank as acontender. In some cases, considerations of space in a complex, compact liquid rocket
vehicle dictate the use of other geometrical shapes (e.g., a torus, or a cone with
hemispherical ends). The design considerations for these special configurations, however, arethe same as those for the more conventional shapes. A disadvantage of the torus and the
cylinder with L/D < 5 (no longitudinal weld) is the weld-length requirement.
Liquid-carrying tanks with positive expulsion devices. - Although the foregoing statements
for liquid-only tanks are applicable to positive expulsion tanks, additional problems inestablishing tank size and shape optimization are introduced by the positive expulsion
device. The design of expulsion device and tank shell are so extensively interdependent that
parallel, simultaneous designs are essentially mandatory. A decision that is made early in the
design effort is the type of expulsion device that will be used. An early decision is necessarybecause of the wide variations in expulsion-device installation requirements. For example, aflexible Teflon bladder can be folded and installed in a tank through a comparatively small
opening (e.g., 4-in. diam.), whereas a corrugated metal diaphragm requires either a bolted
flange or appropriate weld joint at the tank girth. Additional tank-volume deltas introduced
by the expulsion device that must be considered are (1) liquid residuals due to inability ofthe device to expel all the liquid from the tank's liquid compartment, and (2) volume
displacement of the device and associated working clearance required by the device.
Gas-carrying tanks. - With gas pressurants, there is the added problem of significant change
in volume and pressure with temperature. The temperature gradients resulting from heat of
compression during tank charging and decompression cooling during pressurant discharge aswell as from the influence of external environments must be considered so that adequate
strength margins and sufficient volume at time of pressurant demand can be ensured. Tank
pressure charging usually can be programmed to ensure temperature/pressure combinationsthat are consistent with a tank's capabilities. The pressurant consumption schedule, on the
other hand, usually is not known precisely but rather must be predicted for a particular
mission. These hypothetical consumption schedules, necessarily conservative, become the
basis for the depletion analysis. The long lead time required for tank development
frequently forces the tank designer to finalize his design and commit to material and forging
procurement far in advance of refined mission information. When the design involves
comparatively large tanks, the designer must employ keen judgment in finalizing dimensionsand tolerances. For example, a wall thickness tolerance of 0.005 in. on a 40-in.-diam. sphere
constructed of 6Al-4V-tffanium will affect its weight by over 3.5 pounds (4 percent).
15
Becauseof greaterstructuralefficiency,sphericaltanksprovidea weightadvantageoverother configurations. In contrast to liquid-carrying tanks, it is advantageous to design
gas-pressurant tanks to the highest working pressure consistent with the capabilities of the
associated downstream system and thereby attain minimum diameter (or minimum surface
area).
2.2 TANK MATERIAL
Many different material characteristics may be of interest in the design and development of
pressure vessels. However, certain characteristics are of primary importance and can
determine the success or failure of such a project. These characteristics are identified and
considered early in the material evaluation and selection phases of the program. Material
selection usually is based primarily on the following properties:
• Strength/weight efficiency under critical load/temperature conditions (or othercritical failure conditions)
• Fabricability (capability of being fabricated into the desired configurations and
sizes without loss of properties)
• Compatibility with all anticipated environments
• Fracture toughness and resistance to subcritical flaw growth
• Availability of shapes and sizes within required schedules
• Costs of materials and material processing and fabrication.
The most efficient material for tank construction, from the standpoint of load-carrying
ability versus weight, depends upon the type of critical loading, principally whether tension
or compression. Internal pressure usually is the critical load in all tanks that do not form anintegral part of the vehicle structure. These tanks include most upper-stage and
support-system tanks. In such applications, the strength/weight efficiency of candidate
materials can be compared on the basis of the ratio of usable tensile strength to weight.
Usable tensile strength includes provision for the presence of flaws that are of a size below
the limit of reliable NDI detectability or below the size that can be screened by a properly
designed proof test. The vehicle tanks for the first and intermediate stages may be critical
either in compressive buckling due to boost and aerodynamic forces, or in tension due tointernal pressure alone or in combination with structural loads. Materials that are efficient
in compressive buckling generally have a high ratio of elastic modulus to density and a high
ratio of compressive yield strength to density. However, to compare accurately the
16
efficiencyof materialsfor a compressivelyloadedstructure,it isusuallynecessaryto utilizea structuralindexthat representsthe structuralconfigurationandthe loadinganticipated.Thissubjectistreatedin references6 and7.
2.2.1 Mechanical Properties
The mechanical properties of concern in the design and analysis of propulsion-system tanks
are ultimate tensile strength (Ftu), which governs ultimate burst pressure under ductile
failure conditions; tensile yield strength (Fry), because of the requirement that there be no
yielding either at limit load conditions or during proof testing; compressive yield strength
(Fcy) for compression critical structures; and the material elastic properties (E, G, and u).Shear and bearing strength properties (Fsu, Fb ru, and Fb ry) apply to design details such asmechanical attachments and are not normally important factors in the selection of
materials.
High-cycle, low-stress fatigue data sometimes are required to evaluate the effects ofstructural vibration or severe acoustic environment. Low-cycle, high-stress fatigue data often
are used to evaluate the effects of multiple pressurization cycles. The material properties
utilized in fracture-mechanics analyses are discussed in section 2.2.4.
The effects of a number of important variables on mechanical properties must beconsidered. These variables include temperature, thermal exposure, duration of loading,
biaxiality and triaxiality of loading, rate of loading, and unusual environments such ascorrosive fluids and radiation. Design properties are determined for base metal and welds
and sometimes for weld heat-affected zones. The effects of loading direction with respect to
base-metal grain orientation are considered. Properties along the direction of the weld beadas well as across the weld are evaluated. The effects of all processing, forming, and heat
treatments on material design properties are evaluated.
Whenever possible, the precise values of the material mechanical properties used in designand analysis (the "design allowables") are determined by methods that result in consistent
levels of reliability for all materials and conditions of application and service. The military
handbooks, MIL-HDBK-5B (ref. 8) for metals and MIL-HDBK-17A (ref. 9) for plastics,
contain considerable design property data. Nearly all other sources of materials properties
data, unless explicity stated otherwise, contain only typical values that are not suitable per
se as design values.
The methods used to compute design allowable strengths of unflawed materials forMIL-HDBK-5B are covered in MIL-HDBK-5B Guidelines for the Presentation of Data (ref. 8,
ch. 9). Two reliability levels are observed, "A" values, which must be met or exceeded by
the product 99 percent of the time, and "B" values, which must be met or exceeded 90
17
percentof the time. in both values,a statisticalconfidencelevelof 95percentisobserved."A" allowablesareusedfor single-load-pathstructuressuchaspressurevessels.
The determinationof designproperty valuesfor weldspresentssomespecialproblems.Designvaluesarenot currently availablein reference8 for weldedalloysof interestforaerospacepressurevessels.Weldtestdatathatareavailablein the literaturecanbeusedonlyasa guide to the valuesthat might be reliably obtainedin anygivenweldingsetup.Themanyvariablesthat affectweldquality andstrengtharediscussedfurther in section2.2.2.2of this monographand treatedin detail in reference10; methodsfor determiningweldallowablesarealsodiscussedin thesamereference.
The basicapproachfor determiningweld allowablesthat is describedin referencel0 issimilar to the approachrecommendedin the MIL-HDBK-5BGuidelines(ref. 8, ch. 9) formetalsin general(assumingthat processcontrol is exercisedoverall of the significantwelding variables)with but one important exception: the minimum weld strengthdeterminedby statisticalanalysisof test dataon weld couponsmay begivena furtherreduction to account for differencesbetweenthe behaviorof couponsand weldedstructures.Thisreductionfactorhasbeentermeda "coupon/structureratio" (ref. 10,p. 71)and is evaluatedby comparativetestsof couponsandrepresentativestructuressuchassubscaletanks.Valuesof this ratio usedto establishallowableweldstrengthsfor tanksandother structureson the Apollo spacecraftandSaturnS-II stagehavebeenin the rangeof0.80 to 0.90( 10to 20percentreductioninstrength).
2.2.1.1 TEMPERATURE EFFECTS ON PROPERTIES
Reduced temperatures tend to increase material mechanical strength properties, but oftendecrease material ductility and toughness values. Such strength increases sometimes are
utilized in the design and analysis of tanks intended for the containment of cryogens when asignificant weight saving can be realized. However, it is then necessary to ensure that the
fracture toughness of the material is adequate for the anticipated operational and
proof-testing conditions; this subject is treated in detail in section 2.2.4. It is also necessaryto ensure that room-temperature tank pressurizations can be limited to values that are
consistent with the lower room-temperature mechanical properties.
Elevated temperatures tend to reduce material mechanical properties. For some materials
(e.g., titanium alloys), even a small increase in temperature above room temperature results
in a significant reduction in strength (approximately 10 percent at 200 ° F). Temperature
increases on this order can result from compression heating during pressurization ofhigh-pressure gas containers when a proper heat exchanger is not used. If such effects are
not considered, yielding may occur in tanks fabricated from titanium and other alloys thatare sensitive to temperature.
18
Exposureof metals to elevatedtemperaturesfor extendedperiodsof time can causepermanent changes in mechanical strength properties, normally reductions. Thetemperaturesat whichsuchchangesoccurusuallyarenearor abovethematerial'sagingortemperingtemperature.However,exposureof very long duration(monthsor years)canresultin strengthdecreasesat temperatureswell belowthe normalagingtemperatures.
2.2.1.2 FATIGUE STRENGTH
Fluid-containing tanks often are required to withstand a fairly large number of cycles of
pressurization. Such cycles occur during tank acceptance testing, integrated system testing,
tests and operations performed after delivery of spacecraft, and, of course, service
pressurizations. Typically, the total number of such cycles can approach and sometimesexceed one hundred in number. Tank failures have occurred as a result of such repeated
pressurizations. Such low-cycle fatigue failures usually originate at stress-concentrationpoints including preexisting cracks or crack-like flaws. The development (nucleation) of
cracks tends to occur more readily in materials of limited ductility and in locations in which
poor design or fabrication techniques provide localized regions of high stress. Crack
nucleation is avoided by a combination of proper design, material selection, fabrication
techniques, and quality control, and by demonstration of the ability of hardware to meet
cyclic pressurization requirements during qualification testing. The avoidance of failures
resulting from the growth of preexisting cracks during pressure cycles is discussed in section2.2.4.
2.2.1.3 CREEP
Creep is the time-dependent deformation of material under prolonged stress. Pressurization
stresses in tanks tend to be long in duration and high in value with respect to material yield
strengths. Such stress-time histories are likely to cause significant creep in materials at
temperatures for which any creep tendencies have been observed. Although creep is usually
associated with high temperatures, especially as related to the temperature at which
metallurgical processes such as aging or the relief of cold-working effects occur, at least one
notable exception to this rule has been observed: the creep of titanium and titanium alloysat room temperature and at moderately elevated temperatures. The temperature range of
this phenomenon is from a little below room temperature to about 600 ° F, with the
minimum creep resistance (in ratio to material static yield strength) occurring in the regionof 200 ° to 350 ° F. Data on low-temperature creep in titanium alloys are available in
references 11 and 12. The conventional creep of titanium becomes significant at
temperatures above about 750 ° F; however, the behavior of metals at high temperatures
(above about 300 ° F) is considered outside the scope of this monograph.
Aluminum alloys also exhibit a slight tendency to creep at room temperature; however, the
effect is negligible for loading times less than 1000 hours. For longer loading times or for
19
shortertimesat temperaturesabove175° F, creepof aluminumcanbecomesignificant.Data on the creep of aluminum alloys at room and elevated temperatures are provided inreference 13.
2.2.1.4 BIAXIAL-STRESS PROPERTIES
The multiaxial loading of metals can have significant effects on material properties that may
or may not have impact on tank design but are an important consideration in the
interpretation of pressure-test results.
The biaxial tensile stress fields that normally exist in pressurized tanks may improve
material performance, may have no discernible effects, or may even deteriorate
performance. Materials that are ductile, homogeneous, and isotropic may show an increase
in tensile load-carrying ability, the amount depending on the biaxial stress ratio. The
maximum effect usually occurs at a biaxial tensile-stress ratio of 2: 1. In carefully conductedmaterial tests, the magnitude of this effect tends to be on the order of that predicted by the
Von Mises criterion (also termed the octahedral-shear-stress theory and the distortion-energy
theory). According to this theory, the equation for the effective stress for yielding in anelement subjected to a complex stress field is
o.ff = (1/_-) _/(o,, - ay) 2 + (oy - 0.) 2 + (o. - 0,,)2 (1)
where
Oct f = effective normal stress
ax, oy, and Oz = principal stresses, i.e., normal stresses acting on three mutually per-pendicular planes of zero shear stress
Anisotropicmaterials [e.g., "textured" titanium (titanium processed to obtain preferredorientations of the individual crystals or grains)] do not fit equation (1). The effects of
biaxial stresses on such materials and the equation that applies are described in reference 14.
In addition, many homogeneous and nondirectional alloys do not appear to behave in full
accordance with the distortion-energy concept.
The difficulty and cost of accurately evaluating biaxial strengthening effects, together with
the problems associated with correctly applying them, has frequently resulted in these
effects being neglected in the design of liquid rocket propulsion-system tanks. On the other
hand, the existence of such effects should be considered in the evaluation of burst-testresults.
There appears to be a possibility that some materials are reduced in strength when subjectedto biaxial tension. This effect could occur with low-elongation materials in a 1:1 biaxial
20
tensile-stressfield.Thisstressstate(typical of asphericaltank) limits theductility effectstotheplanethat includesthematerialthicknessdimension.A materialhavinglimitedductilityin thethicknessdirectionwill tendto behavein abrittle fashionin suchastressfield.
The questionasto the possibilityof decreasesin materialresistanceto crackpropagationdue to biaxialtensilestressesappearsto remainunresolved.Datareportedin reference15appearto indicatethe existenceof suchreductions.Until the effectsof biaxialloadingonthe propertiesof materialsregardinggrowthof critical andsubcriticalflawsaremorefullyestablished,it shouldbeassumedthat biaxialloadingmaydecreaseandprobablywill notbenefittheseproperties.
2.2.2 Fabrication Considerations
Fabricability is one of the determining characteristics in the choice of materials for
aerospace vehicle pressure vessels. Not all high-strength structural materials can be fabricatedeconomically into reliable vessels. The following essential requirements must be satisfied:
(1) The material must be available in suitable forms, sizes, and levels of quality within
the necessary schedules.
(2) The material must be capable of being formed and machined to the required
configurations, on the available equipment, and at the material thicknesses andstrength levels determined to be appropriate.
(3) The material (alloy) must possess sufficient weldability to suit proposed methods
of assembly that involve welding.
(4) The material thermal processing requirements must be capable of being met,
within existing economic restrictions, on actual parts or assemblies for which
thermal processing will be necessary.
2.2.2.1 SHAPING AND FORMING
Methods of fabrication that have been used to produce metal pressure vessels for aerospace
vehicles are compared in table IV (adapted from ref. 16). The factors of size, shape, and
material formability have considerable influence on the suitability of desired material and
on the methods of fabrication and heat treatment that can be employed. Large size in
particular often limits the selection of fabrication methods and heat treatments. Suchrestrictions may have impact on the optimum material selections. Also, large size often
causes the relief of weld residual stresses in completed vessels to be economically if not
21
Table IV. - Comparison of Fabrication Methods for Pressure-Vessel Com
Component
End domes
(complete
heads)
Cylinders
Fabrication method
Drawing:
Hydropress (trapped rubber
forming)
Hydroform (hydraulic fluid
forming)
Advantages
High.Energy.Rate Forming:
Explosive
Electrical (including spark
discharge and magneto-
dynamics)
Conventional
(manual or power)
Spinning:
Shear
Forging
Segmenting (formed and
welded segments)
Roiling and Welding
Shear spinning
Moderate production rateModerate tooling costs
Larger sizes than hydroform
High production rate
Better thickness control than hydropress
Very large potential sizes (depending onavailable facility)
Good reproducibility
Low to moderate tooling costs
High production rate
Good reproducibility
Permits integral bosses and skirtsCan handle thick material
Good thickness control
Spinning can be performed hot
Moderately large sizes
Low tooling oosts
Not limited to materials with cold- or warm-
forming ability
Permits complex configurations
Permits integral attachments
Large size capability (starting with smaller
individual parts)
Reduces difficulty and cost of forming
Accommodates large sizes
Low cost, simple process
Eliminates longitudinal welds
Permits integral reinforcements
Provides good thickness controlForming can be performed hot
,onents (adapted from ref. 16)
Disadvantages
Part size and thickness limited
Temperature broiledPoor control of thickness
Limited to small sizes
Temperature limitedRelatively high tooling costs
Limited to cold fornung
Low production rateLimited availability of facilities
Limited to small sizes
Requires specialized equipment and tooling
Size limited
Limited availability of equipment
Poor thickness control
Permits no integral details as formed
Temperature limitedThickness limited
Low production rate
Requires ductile material
Size limited
High costs
Requires considerable machining
Low production rate
High total costs - toohng, welding, and
inspectionPotential for reduced reliability due to
increased weldingPoor dimensional control
Very low production rate
Potential for reduced reliability due to
longitudinal weldPermits no integral reinforcements as
fabricated
High cost for low production quantity
Limited equipment availabilitySome limitations on size
22
technicallyimpractical,especiallywherethe numberof units to be manufacturedis toosmall to justify a largeinvestmentin constructionof specialheat-treatmentfacilities.Materialschosenfor such tanksare requiredto havegoodpropertiesand toughnessinas-weldedwelds.The impact of heat-treatmentrequirementson material selectionisdiscussedfurther insection2.2.2.3.
Detailedinformationondeformationprocessingof variousaluminum,titanium,iron,nickel,andcobaltalloysisprovidedin references17through23.
2.2.2.2 WELDING
Material welding characteristics and weld properties are primary considerations in theselection of metals for tanks. A weldable material is one that can be fused without the
formation of deleterious phases or constituents either in the fusion zone or in adjacent
heated areas, has sufficient ductility (both in the bead and in adjacent areas) from the
melting temperature to room temperature to resist cracking, and has suitable strength andfracture resistance either as-welded or on completion of postweld thermal processing. From
the standpoint of practicality, a weldable material should also be amenable to repair welding
procedures without inherent tendencies toward the formation of new defects or significant
impairment of properties.
The strength of weld metal usually is lower than that of parent metal. This difference can
result from a variety of causes, as follows:
• Welding alloys that obtain their strength by cold working, which cannot be
performed after welding
• Filler alloys that have good ductility but lower strength than the parent metal, or
that lack heat-treat response
• Lack of proper heat treatment after welding
• Lack of strength in the cast-weld deposits, even after heat treatment
• Defects both in weld metal and in weld geometry
Increased scatter in weld properties as compared with parent metal, the result
being similar average strengths but lower statistically computed design strengthsfor the welds.
Reduced design strengths in welds usually are compensated for by providing extra thickness
in weld joint areas. Details of weld joint designs are provided in section 2.3.6.1.
23
Weldsusuallyare lessductile than parent metal (although occasionally the reverse is truebecause of the use of ductile filler material or lack of heat treatment in the weld area) and
therefore usually are less able to withstand the effects of stress or strain concentration. For
this reason, welds generally are located away from areas such as abrupt changes in contour,
where high local strains are likely to occur as the tank is pressurized. For the same reason,stress concentrations in the weld itself (e.g., at large unmachined beads or other geometric
irregularities) are undesirable.
In the as-welded condition, welds contain residual stresses unless special welding processes
or operations designed to prevent them are used. In general, the severity of such residualstresses increases with increasing material thickness and number of weld passes, low thermal
conductivity, high thermal expansion, and high modulus of elasticity. On the other hand,
processes such as pressure welding and forge welding may result in no significant residualstress.
It is often necessary to relieve residual stresses in tanks to prevent cracking, warpage,
reduced fatigue strength, or reduced reliability against fracture in general. Tank welds
usually are stress relieved during subsequent thermal processing such as aging of titaniumand heat treatment of low-alloy steel. Welds in the as-welded condition have been successful
in some tanks (e.g., those in large boosters or lower stages). The alloys used in these
applications have relatively high ductility and good fracture toughness in the welds andadjacent material; such alloys include 2014, 2219, 6061, and 5000-series aluminum alloys,and the AISI 300-series stainless steels.
Detailed information on welding can be found in references 24 through 26.
2.2.2.3 THERMAL PROCESSING
Thermal processing procedures for alloys of interest for high-strength pressure vessels are sodetailed and variable that it is not feasible to attempt more than an outline of this subject in
this monograph, emphasis being given to those details that have possible impact on designand material selection. Detailed information on heat treatment can be found in references
27 through 39, or obtained from the producers of specific alloys.
Before an alloy can be selected for a proposed tank, it is necessary to determine if the
thermal processing required to obtain the desired design properties can be accommodatedwithin an economically feasible manufacturing plan. Depending on material forming,
machining, and welding characteristics, on the need for postweld thermal treatment to
relieve residual stresses or increase weld strength, and on the difficulty of heat treating
completed components or tanks due to size, complexity, possibility of distortion, and
similar considerations, thermal processing may be performed in the following phases ofmanufacture:
• As received or prior to any fabrication
24
• Duringor afterfabrication
• After roughmachining
• Onweldedcomponentsor completedtanks.
The important characteristicsof thermalprocessesapplicableto alloysof greatestcurrentinterestfor high-strengthpressurevesselsareoutlinedin tableV. (This tableis intendedtoassistin materialselectionstudiesratherthanto providedetailedprocessinginformationonspecificalloys).
Annealingfrequently is requiredprior to or during forming of tank components,eitherbecauseof limitations on materialductility or limitations associatedwith the formingprocessor capacity(power)of availableequipment.The annealingprocessindicatedmayeither be a full anneal,as shownin table V, or a lower temperature"processanneal"(processannealingtreatmentswere omitted from table V for the sakeof simplicity).However,a full annealingtreatmentnormallycannotbeperformedonanalloythat obtainsstrengthfrom cold working,sincethe strengthof suchmaterialscannotusuallyberestoredby anyprocessthatcanbeperformedoncompletedtankcomponents.
Severalof the alloys listed in table V have considerableformability in hardenedorpartially-hardenedconditions.Theseincludethe aluminumalloysshown(especiallywhennot work hardenedor agedto thehigheststrengthlevels);HY140,9 nickel- 4 cobalt,and18-nickelmaragingsteels;andwork-hardenedausteniticstainlesssteels(exceptingfull hardandhardertempers).
Manyof thematerialslistedin tableV donot requireheattreatmentafterwelding,eithertorestoremechanicalpropertiesor to relieveresidualstress.Alloys thathavebeenusedin theas-weldedcondition or that havepotential for suchare shownwith a checkmark in thecolumn "None", under the generalheading,"Postweldthermalprocessesperformedorrequired".Someof thesealloysalsohavecheckmarksin othercolumns,indicatingthat itmaybenecessaryin someapplicationsto thermallyprocessthe materialafterwelding.Anexampleof thelattercategoryis 18-percent-nickelmaragingsteel,whichhashadlimitedusein the as-weldedconditionbut which requiresagingafter weldingto developmaximummechanicalpropertiesin welds.
Aluminumalloysthat must be formedin the as-solution-treatedconditionaresometimesagedafter forming and are not further heat treated after welding. However,in thefabricationof largeparts,agingmayresultin growththat is both significantandvariable,andtherebyproducedifficultiesin fitup of partsfor welding.Thisproblemoccurredduringfabricationof the large2014-aluminumcylindersfor theSaturnS-IIstage.It wassolvedbycontouringcylindricalsectionsin theT6 condition.
25
I
0
¢s
mss
E
u
0. x ;.: x
:,.<
X
X X
o-a
m
v
._®
m • u
i _ 0_ ,,
o o_
NNNNNuuu_u u
oo o_ ,_° oo °o °o o _ o_ _ oo
o . o° _ _ _ ._ ,._ ,._ ,-_ ,-" i_ i._ ,._ _ ._ ._ .:
kl ._.o Z •
>.E0
o
26
X X
_._ o
X X
A
>,
=.,
_o _o
. _;
o, _- m
.
f
e _
_ g
u =
u
27
Dimensional changes occur during heat treatment or aging of many other alloys of interest
for tank construction, particularly in phase-change materials such as alloy steels and
precipitation-hardening stainless steels. This factor has affected manufacturing planning and
even the suitability of alloys for proposed designs.
Residual stresses that are introduced into titanium alloys and most martensitic steels (except
as noted in Table V) during welding normally require thermal relief. In the case of
heat-treatable titanium alloys, relief is often accomplished in a combined stress relief and
aging treatment after welding. Conventional low-alloy steels must usually be fully heat
treated after welding, and weld stress relief takes place during this process.
Annealing treatments to remove residual stresses resulting either from deformation
processing or from welding vary appreciably according to the degree of stress relief sought,
the temperatures at which loss of material hardening begins (for alloys heat treated prior to
forming or welding), and on practical difficulties in heating components or assemblies. It is
usually necessary to tailor such processes to specific materials and applications, and suchdetailed information is beyond the scope of this monograph.
Table V indicates the types of alloys that require high temperatures, special atmospheres, or
rapid quenching during heat-treatment processes preparatory to hardening. Such processes
are difficult and costly to perform on completed components, especially in large sizes.
Special furnaces, fixturing, handling, and quenching facilities that may be required may notbe economically justified by the number of tanks to be fabricated.
2.2.3 Material Compatibility with Environments
Materials selected for liquid rocket system tanks must be compatible with the fluids to becontained. Alone or in combination with a suitable protective finish, tank materials must
also be resistant to the effects of exposure to all external environments encountered. Inaddition, the contamination or deterioration of tank materials during material processing,
manufacturing, inspection, test, transportation, and storage must be prevented by avoiding
exposure to or providing protection from all fluids or processes that are known to have
deleterious effects. The undesirable metal/fluid reactions, the possible consequences, and
the major sources of such reactions are summarized in table VI.
Metal/fluid incompatibility reactions are discussed in the following sections, first from the
standpoint of the major sources of reactions that must be avoided, and second from the
standpoint of specific failure mechanisms and their avoidance.
28
Table VI. - Causes and Effects of Metal/Fluid Reactions
Metal/Fluidreaction
Metal corrosion
(including general
corrosion, pitting,
intergranular corrosion,and chemical attack)
Possible consequences
to system
(1) Metal weakening throughloss in cross.sectional
area and introductionof stress raisers
(2) System contaminationwith corrosion products
A
X
Catalytic decomposition Loss of efficiency or contamin-
of propellants ation of system or both
Hydrogen embrittlement Brittle fracture at low Xof steel stresses, especially under
long.duration loading
Contamination of titanium Brittle fracture at low stresses X
Stress corrosion Metal crack growth or fracture Xat reduced stress levels
Galvanic corrosion (1) Rapid deterioration of material X(2) Stress-corrosion failure
Hydrogen-environment Embrittled behavior of metal whileembrittlement of metals exposed to hydrogen gas
Ignition of materials Catastrophic combustion
Major sources ofmetal/fluid
reaction
B C
X X
X
D
X
X X X
X X X
X X X
X
X
Notes:
A=
B =C=
D=
manufacturing fluids and processes
proof and system testingservice fluid containment
atmospheric exposure
29
2.2.3.1 SOURCES OF MATERIAL/FLUID REACTIONS
2.2.3.1.1 Manufacturing Fluids and Processes
Many high-strength alloys arc susceptible to attack or contamination by rather Coilnnonly
used manufacturing fluids and processes. Some of the potential sources of such reactions are
lubricants; cleaning agents, solvents, or baths: etching, descaling, stripping, or brightening
solutions; identification marking materials: nondestructive inspection fluids: electrolytic and
electroless plating operations: electrolytic metal removal or cleaning processes: chemical
milling; electrical discharge machining: heat treatment and welding. These materials and
processes are not to be condemned per me: however, they have been sources of undesirable
reactions and may require verification of compatibility before being used in a new
application.
Of the alloys of interest for tank constrtlction, titanium alloys are the most susceptible to
contamination. Most of the categories of materials and processes listed above are potential
sources of trouble for titanium. The most frequently occurring mechanisms of titanium
contamination during nmnufacturing operations are as follows:
• Hydrogen contalnination from room-temperature processes or lrom contact with
hydrogen (or gases that contain the element hydrogen) at high temperatures
• Oxygen and nitrogen contamination from contact with these gases (as in air) at
high temperatures
• Halogen contamination from the use of halide-containing materials on titanium
prior to heat treatment or welding.
Hydrogen contamination embrittles titanium when concentrations exceed about 100 to 150
parts per million, the effective concentration depending on tile type of alloy; the subject of
hydrogen in titanitun is treated in references 31 and 32. Oxygen and nitrogen also tend to
embrittle titanium. The depth of enlbrittlement resulting from heat treatment in air is such
that it is often removable by machining, ttowever, when titanium alloys are welded in air,
the entire weld is embrittled. The subject of titanium oxidation and contamination is
treated in reference 33. A listing of materials and processes found to be compatible with
titanium alloys is provided in reference 34, together with a list of substances that have
proven to be incompatible.
The embrittlement of steel by internal contamination with hydrogen has been well
documented: and current procedures for production, processing, and manufacturing of steel
reflect the need to prevent embrittlenaent by avoiding the sources of such contamination or
by removing the contaminant if it cannot be avoided. Typical sources of contamination are
electrolytic processes such as electroplating and cathodic cleaning; electroless plating,
3O
pickling, and acid-stripping baths: chemical milling: and heat treating or welding in
atmospheres contaminated with hydrogen or water.
The susceptibility of steels to hydrogen embrittlement increases with increasing strength
level. Iligh-strength, low-alloy steels are the most susceptible to the effects of hydrogen:
however, it may be assumed that all steels of interest in the present context excepting
austenitic stainless steel IAISI 300 series) are to some degree affected. In reference 35 (a
user specification for the prevention and elimination of hydrogen embrittlement in steel),
process control is indicated for all alloy steels over 140 ksi (tensile strength), all
precipitation-hardening stainless steels above 160 ksi, all martensitic stainless steels above
180 ksi, case-hardening steels, carbon and alloy spring steels, and tool steels. The subject of
embrittlement of steel by hydrogen contamination is treated in considerable detail in
references 36, 37, and 38.
Other types of alloys of interest for liquid rocket tanks and components, viz., alunlinunl
alloys and nickel- and cobalt-base high-temperature alloys, are not susceptible to
hydrogen-contamination embrittlement. However, such alloys may exhibit a surface-related
brittleness during exposure to gaseous hydrogen at temperatures near ambient temperature,
as discussed in section 2.2.3.2.3. This latter phenomenon is distinguished from the
embrittlement of titanium and steel via internal contamination.
2.2.3.1.2 Testing Fluids
Fluids used in tank and system testing are a potential source of reactions with tank
materials: these fluids therefore nmst be properly evaluated before use and properly
controlled in use. Although such fluids usually are not corrosive in nature, their formulation
must be properly chosen and any contamination avoided so that local attack that can
nucleate fracture at a lowered pressure may be prevented. However, the most noteworthy
consequences of incompatible testing fluids have been instances of stress corrosion (sec.
2.2.3.2.1) or reduced threshold stress intensities for crack growth (sec. 2.2.4.1). Examples
are tile premature failure of titanium tanks pressurized with methanol and of steel tanks
presst|rized with water (refs. 39 and 40, resp.).
2.2.3.1.3 Stored Fluids
Liquid propellants, particularly oxidizers, often possess a highly aggressive or reactivechemical nature. Tile containment of such substances with metals that also may be quite
reactive often is possible only because of the protective mechanism of the fornmlation of a
relatively stable oxide fihn or layer on the metal surface. Because of the inherent potential
reactivity, it is often difficult to predict the circumstances under which metals will resist
such aggressive fluids or, alternatively, will be rapidly attacked, or will fail from stress
corrosion or some other failure mechanisln. For this reason, it has been found necessary to
verify the compatibility of proposed metal/fluid combinations under conditions that
31
representthoseanticipatedin service,particularly tile Factorsof ctlemicalcomposition(inchtdingsuspectedFluidcontanlinants),tcmpcratttre,anddurationof exposure.Exposurestressis alsoin3poltanl,asindicatedinsection2.2.3.2.1.l'lxposurcpressurenlayor maynotbe importanl.The effectsof afluid on crackgrowthshouldalsobeconsidered,asdiscussedin section2.2.4.1.
Besidescoml_osition,the materials-orientedFactorsafFectingthe rateof attackor degreeofincompatibility includeheat-treatcondition,existenceof cold work or residualstressesatexposedsurl'aces,and exposure oF lnetal end grain _which oltetl results in much greater
attack than exposure of the other grain dimensions). Also, welds may be affected in an
entirely diFFerenl manner From the parent (unwelded) metal because of changes in chemistry
and structt_re during welding.
Propellant decomposition may be accelerated by tile catalytic effects of certain metals. In
tile selection of materials, this possibility usually is investigated, particularly for metals used
in contact with hydrogen peroxide and hydrazine-type propelhmts.
('onsiderable data on tile compatibility of various metals and nonmetals with a number of
different liquid rocket fuels and oxidizers are available m references 42 and 43. Reference
43 contains a summary of data on corrosive attack rates, the occurrence of propellant
catalytic decomposition, and impact ignition. Additional discussion ol" material Failure
mechanisms restllting l'ronl reactions with stored lluids is contained in section 2.2.3.2.
2.2.3.1.4 Atmospheric and Environmental Corrosion
Integral tanks that Form a portion of tile exterior surface of a vehicle usually are subjected
to direct atmospheric exposure during a portion of tile vehicle life. However, vessels that are
not so exposed after system installation in the vehicle usually are exposed indirectly to
corrosive atmospheric conditions, principally moisture-laden air with or without the
contributing effects of salt or industrial chemicals. Such exposure occurs during
nmnulacturing, storage, testing, transportation, and vehicle operation.
Some alloys are quite resistant to atmospheric corrosion because of the formation of a thin,
tightly adherent, protective oxide fihn. Titaniunl: the high-nickel, high-chromium stainless
steels and superalloys: and some aluminunl alloys are in tllis category. However, really other
high-slrength alloys of interest for liquid rocket tanks and associated hardware (e.g.,
2000-series aluminum alloys and low-alloy steels) must be provided with a protective finish.
Re|erences 43, 44, and 45 often are used as sources of information on tile selection of metal
finishes For spacecra|t hardware. Table 11 of reference 43 provides a convenient guide to the
alloys that normally need a Finish For corrosion protection, ttowever, specific space vehicles
usually have requirements that are likely to diFFer from those envisioned in such general
specifications. Such requirements may include operation in a hard vacutlm, thermal control
in space, and identification, for example. For this reason, Finish requirements are usually
32
defined and documentedfor specific spaccvehicles.Examplesof this are the finishspecificationspreparedfor tile ApollospacecraftandtheSaturnS-IIstage(refs.4(_and47,resp.).
Thechoiceof nlaterialsandfinisheslnUslalsoconsiderthepossibilityof stresscorrc_sJollor
galvanic corrosion in atmospheric environnaents. These faihlre mechanisms are treated in the
following sections.
2.2.3.2 TYPES OF MATERIAL/FLUID REACTIONS
This section treats failure mechanisms involving material reactions with fluid envil-onments
encountered in service applications: these phenomena nlust be considered in the design and
selection of materials for liquid propulsion systems.
2.2.3.2.1 Stress-Corrosion Cracking
Stress-corrosion cracking is onc of the more conlmon sources of failure of highly stressed
metals. This form of cracking nlay be defined as delayed fracture rcsulting from the
colllbined action of stress and a corrosive environment. Since stress corrosion can occur in
environments that produce little or no ordinary corrosion, the term "aggressive
environment" often is used to denote environments that react unfavorably with some metals
or are suspected of doing so. Corrosive or aggressive environments to which liquid
propulsion system tanks and associated hardware are exposed may be either a fluid
contained in the tank {during testing or in service} or an external environment. The most
common external environment is ambient air containing moisture and traces of salt and
other chemicals.
The occurrence of stress corrosion depends on three basic factors: the severity of the
environment for the particular material, the severity of the sustained stress, and the duration
of exposure. Therefore, avoiding this type of failure requires ( 1 ) avoiding metal exposure to
aggressive environments or selecting materials that are resistant to such environments, (2)
minimizing the severity of sustained tensile stresses at exposed metal surfaces, and (3) (if
possible) minimizing exposure duration.
The subject of stress-corrosion cracking and its prevention (especially in military equipment')
is treated in military handbook MIL-HDBK-724 (ref. 48): table i in this reference provides a
convenient listing of alloys and associated environments for which stress-corrosion cracking
has been reported. Stress-corrosion data on a variety of alloys and environments may be
found in reference 49; table 1.1 of this reference also contains a relatively comprehensive
listing of alloy families and corrosive materials for which stress-corrosion cracking has been
observed. In addition to these two sources of data, reference 50 is a valuable collection of
data on stress corrosion of titanium. This reference covers the corrosive effects of molten
33
alloys such as cadmiunl and also the reaction between solid silver and silver chloride in
contact with titaniunl at temperatures above or about 600 ° F. A more complete listing of
low-melting-point metals that may not be used in contact with titanium at elevated
temperatures is provided in reference 34 (table VIII.
Titanium stress corrosion with nitric oxide (NO)-free N204 and also with methyl alcohol
during the Apollo program are documented in reference 39. This reference also contains the
results of compatibility testing of many other materials in contact with titanium. The
titanium/N2 O4 incompatibility problem was solved for the Apollo program by maintaining a
sufficient NO content to inhibit the reaction between titanium and N204 . Specification
NASA MSC PP1)-2 for propellant grade N204 (ref. 51_ currently specifies NO content at
0.60 to 1.00 percent.
The incompatibility of titanium and titanium alloys with the fluorocarbon solvent Freon
MF ttrichlorofluoromcthane_ is treated in reference 52. Indications of titanium
incompatibility with the products of a reaction between small amounts of the solvent Freon
TF {trifluorotrichloromethane} and hydrazine (N 2 tl 4 ) are discussed in reference 53.
Many high-strength alloys of interest for tank construction are subject to stress-corrosion
cracking in atmospheric environments. Many aluminum alloys in the 2000 and 7000 series
are susceptible, some at quite low levels of sustained stress, the response depending on
material heat-treat condition and direction of the stress with respect to the material grain
direction. Conventional low-alloy steels are also susceptible, especialy when heat treated to
high strength levels. Precipitation-hardening stainless steels vary, as a group, in resistance to
this failure mode. Some of these materials are quite sensitive and others quite resistant.
Titanium alloys normally resist atmospheric effects, but are susceptible to salt and other
chlorine sources that rclnain in contact with the metal at temperatures above about 550 ° F.
Some titanium alloys have also been found to be sensitive to sea water at room
temperatures. A comparison of the resistance of various alloys to stress corrosion in
atmosphere and other natural or simulated natural environments is provided in table VI I.
The use of protective finishes and coatings to prevent stress corrosion failures of metals
during exposure to natural environments generally has not met with success when there
existed a pronounced susceptibility to this failure mode. This has been indicated by service
experience with alunfinum alloys 7075-T6 and 7079-T6 and by tests performed to ewduate
the effectiveness of various coatings for the protection of high-strength steel from stresscorrosion {ref. 541.
Currently two distinct approaches are being used to evaluate the effects of stress in corrosive
environments. One might be called tile smooth-specimen or conventional-testing approach
and the other, the fracture-mechanics approach. Conventional testing techniques, which
involve exposure of specimens stressed either in bending or in direct tension, are described
in refercnce 4 _) (oh. 12). The fracture-mechanics approach involves exposing precracked
34
35
_!=
,,I
o c£
o "_ -_
L
iI
>.._=
F I
_i _- ,
i
i
_ _-_ _ ,_
,c
.q.
"5
_.--=_ ×
.,_o
._y ..... .,
-_=o__. " -" Z
_ _, _' _ =
_=
36
specimens to the test environment. In such tests, tile load level and the crack length
combine to produce a local stress intensity that must be less than that which causes crack
growth in dry air (or other nonreactive reference environment) but still high enough to
reveal any environment-caused promotion of crack growth. This method of testing is
discussed in section 2.2.4.1 and is described in detail in reference 63.
The fracture-mechanics approach often is a much more sensitive indicator of
stress/environment effects; it can show significant reductions in material resistance to crack
growth in media for which no stress-corrosion effect is observed in conventional tests.
Values for threshold stress intensity are also required for the implementation of a
fracture-mechanics-based safe-life analysis as discussed in section 2.2.4.2 of this monograph.
2.2.3.2.2 Galvanic Corrosion
The exposure of electrically connected dissimilar metals to an electrolyte results in
generation of an electric current and rapid attack on the "less noble" of the two metals. The
metal dissimilarity that provides the driving force for such reactions is dissimilarity in the
electrode potential developed by each of the two metals when in contact with the fluid. The
circuit must be completed by the electrical conductivity of the fluid in which the two
metals are immersed and also by electrical contact between the two metals or electrodes.
(The latter path is analogous to "shorting out" the terminals of a battery.) The greater the
electrode potential difference, and the more conductive the current paths, the more rapid
will be the attack.
Galvanic action is present in many corrosion processes that do not appear to involve
dissimilar metals. In such cases, the dissimilarity may be between metal phases in the
microstructure or between metal grain boundary areas and the metal grain, or it may be a
concentration gradient in the solution, which results in different electrode potentials in
different areas on the same piece of metal. However, these processes are il3cluded under the
category of corrosion, and the present concern is with the gross form of galvanic corrosion
that results from improper use of dissimilar metals.
Galvanic corrosion can occur within a system in which an electrically conductive fluid is
stored: it may occur outside the system as a result of the effects of atmospheric moisture.
Within a system, the distances spanned by such effects can be as long as the electrical paths
provided by the fluids on the one hand and the metal circuit on the other. Metal
dissimilarities may occur between the tank material and metals used for internal design
details such as slosh baffles, reservoirs, filters, and screens. Exterior to the system, galvanic
action usually is limited to metals that are in contact, in very close proximity, or in the same
moisture trap, if such exists.
Within liquid propellant systems, corrosion of dissimilar metals frequently is of little or no
concern because most propellants either have little electrical conductivity or do not develop
37
significantelectrodepotentialsin contactwith normalstructuralmetalsor both.However,the possibility exists that new or inadequatelytested propellantswill be capableofSUl_portingor promotinggalvanicactioneitheraloneor togetherwith contaminants.Also,galvaniccorrosioncan retidilyoccurwithin fluid propulsionsystemsduringproceduresinwhich liquids other thanpropellants(particularlywater)are in contactwith galvanicallydissimilarmetals.
The permissibleand undesirablecombinationsof metalsfrom the standpointof galvaniccorrosionin the atmosphereor in aqueoussolutionsaregivenin references44 and45, tileanodicmemberof eachcombinationalsois indicated.Additionalandmoredetaileddataonthe susceptibilityof dissimilarmetalsto galvaniccorrosionin solutionsof varyingpH areprovidedin reference64.(It shouldbenotedthat galvanicelectrodepotentialsdeterminedusingaqueoussohitionsarenotvalid for nonaqueoussolutions.}
Protectivecoatingssometimesareusedto preventgalvanicactionwhenit is desiredto risedissimilarmetalsin a way that mightbeconduciveto galvaniccorrosion,ttowever,coatingsimproperly usedcan worsenthe situation. For instance,when the anodicmemberof adissimilarpair is the only membercoated,a smallcoatingdefect restiltsin a very smallanodefacinga largecathode;rapidattackresultsfrom alow ratioof anode-to-cathodearea.Also, platingsthat arecathodicwith respectto the substratemetalcancauseacceleratedattacklit localplatingdefects.
2.2.3.2.3 Hydrogen-Environment Embrittlement
It has been observed recently that ['or many metals the tensile strength, notched tensile
strength, fatigue strength, resistance to crack growth, and ductility are decreased, sometimes
seriously, when the metal is tested in a hydrogen-gas environment. This phenomenon differs
significantly from the phenomena of hydrogen-contamination embrittlement of steel and of
titanium alloys discussed in section 2.2.3. ]. 1. Embrittlement in hydrogen-gas environment
appears to be a surface-related phenomenon that (with the possible exception of titanitun
alloysl persists only while the material surface actually is exposed to the hydrogen gas. A
kirge number of different metals have shown such effects, and, in fact, only a few materials
of interest for tank construction have been found to be essentially free from this
phenonlenon.
The reductions in strength, toughness, and ductility of metals have occurred during
hydrogen exposure both at room temperature and at moderately depressed and moderately
elevated temperatures. The effects have been observed at one atmosphere of presstire, but
are more noticeable with increased gas pressure. Tile greatest effects have occurred in metals
that had been heat treated to maximunl strength levels, ttigh-purity hydrogen normally is
used in laboratory investigations of hydrogen embrittlement; however, the number of
instances of rapid crack growth in steel hydrogen-storage systems now attributed to this
embrittlenlent indicates that it is a real problem for systems that contain gaseous hydrogen.
38
Thisphenomenonis treatedin detail in references¢75through68; considerableusefuldataareprovidedthereinonalloysot"interestfor lank construction.Serviceexperience,inwhichrapidgrowthof cracksin low-carbonsteeland400-seriesstainlesssteelwasencounteredingaseous-hydrogenstoragesystems,isdescribedin reference60.
Titaniumandtitaniunlalloysappearto beaspecialcaseof tile effectsof hydrogenexposureat room temperatures,lixposure to hydrogengascan produceboth the environmentaleffects describedabove that do not persistafter the exposureand, with exposureofsufficientlylongduration,elnbrittlementthatgoesbeyondthesurfaceandremainsaftertileexposure.Thelatter mcchanisnlof embrittlemelltactuallyisoneof internalcontaminationand formation of hydrides.No other alloy systemof interestfor tank constructionhasshownanyevidenceof internalcontanlinationresultingfrom roonl-temperatureexposuretohydrogengas.Titaniunl contaminationwith hydrogengasat temperaturesnear roomtemperatureis treatedin references70 through73.
2.2.3.2.4 Material Ignition
The susceptibility of titanium to potentially catastrophic ignition under conditions of
impact or other instances of localized high-energy pulses in the presence of strong oxidizers
now is widely known. This phenomenon was first observed in red fuming nitric acid
(RFNA) and liquid oxygen (LOX). Reactions also have been observed in liquid lluorine and
in mixtures of liquid fluorine and liquid oxygen (FLOX), and in pure gaseous oxygen at
pressures on tile order of four atnlospheres and greater. Nitrogen tetroxide (N= 04 ) has also
shown a tendency toward such ignition reactions with titanium, but such reactions have
been nonpropagating in nature. (The more serious problenl of stress corrosion of titanium
when it is exposed to nitrous oxide- free N204 was discussed briefy in section 2.2.3.2.1.)
Tile reactions observed with liquid fluorine have also been nonpropagating rather than
catastropic in nature. Further information on titaniunl pyrophoric reactions with fuming
nitric acid, liquid oxygen, and gaseous oxygen may be found in references 74 through 77.
Tile method for deternlining material sensitivity to impact ignition usually involves a
drop-weight device such as tile ABMA apparatus indicated in reference 78. There should be
no evidence of a reaction during twenty successive impacts of a hardened steel striker, each
strike at 72 ft-lbf impact energy. Indications of a reaction are an audible explosion, a visible
flash in a darkened room, or post-test evidence of discoloration or burning.
Additional information on LOX compatibility testing procedures may be found in reference
79. A large nunlber of nonmetallic materials, particularly organic materials, also arc
incompatible with strong oxidizers because of the possibility of ignition. Impact may or
may not be required to set off such reactions.
Undesirable or hazardous reactions also nlay occur between propellants in the hydrazine
family and copper, lead, zinc, molybdenum (and alloys having a significant molybdenum
content), and plain carbon and low-alloy steel.
39
The above instances do not cover this area of concern completely. The compatibility of each
new material/propellant combination must be determined. Considerable data on ignition
and other incompatibility reactions between various materials and propellants are availablein references 41 and 42.
2.2.4 Fracture Control
The rapid development of linear-elastic fracture mechanics has provided pressure-vessel
designers with a new tool to help solve the problem of unexpected brittle failures at stress
levels less than the material yield strength. Previous methods for evaluating material
resistance to brittle failures, such as Charpy impact strength or notched/unnotched tensile
strength, primarily are used qualitatively, whereas the fracture-mechanics approach permitsquantitative evaluations.
This approach, based on the fact that flaws of various sizes exist in all materials, requiredthe development of a mathematical model for the growth of such flaws under stress in
materials that exhibit both elastic and plastic behavior. This mathematical model treats the
flaw as a crack and provides quantitative relationships among the crack dimensions, the
applied gross section stress, and the stress intensity at the tip of the crack. Extension of the
crack is dependent upon the stress intensity at its tip. The stress intensity factor, K, around
the perimeter of a buried elliptical sharp crack in an infinite elastic solid under uniform
normal tensile stress, o, is described by the expression
K = o _ -g {sin2 /_+ (a/c)2 cos 2 _,a (2)
where/3 (the angle designating the location of K along the crack front), a, and c are definedas shown in figure 3 (ref. 3).
The shape factor _ is expressed by
fo hI2_P= ql -k 2 sin = 0'd0; k=_-a2/c 2 forc>a (3)
Values of 6, the complete elliptic integral of the second kind, are published in mathematical
handbooks. A very useful expression developed by Rawe (ref. 801 for approximating q_2 is
_2 = 1 + 4.593 (a/2c) 1"6s (4)
This expression is known to be accurate for flaw aspect ratios (a/2c) between 0.05 and 0.5.
Irwin (ref. 81) adapted the expression given by equation (2) to the case of the part-throughsurface crack under uniform tensile stress by applying a multiplying factor of about 1.1.
4O
o
Cr c<o nn0 Idisplacement
Thickness
FIt
L
I
D t
Back surface ,,/t_
.]-T
Figure 3. - Sketch illustrating surface-crack model and nomenclatureused in fracture-mechanics analysis (ref. 3).
This factor was originally meant to represent the combined effects of both the front and
back free surfaces being made free from normal shear stresses. For the surface flaw in the
0 odepth direction (/3 = 0 ) equation (2) then can be rewritten as
K= 1.1 a _ 5 (5)
Irwin also proposed adding an estimate of tile size of the plastic zone ahead of the crack tip
to the crack size. Adding the estimated plastic zone radius ry to the crack depth a gives, for
the surface flaw in the depth direction,
1.1o
K 4_ #71"(2 + ry )
For plane-strain conditions ry is estimated from
1ry = _ (K/oys) 2
,,x/zTr
(6)
(7)
where Oy_ is material yield strength.
Substituting equation (7) into equation (6) and solving for K then gives
K= 1.1 o x,/'ff-fi/#4) 2 - 0.212 (o/Oys) 2(8)
41
Theexpressionundertile radical in tile denominator is identified as Q (the Irwin flaw shape
and plasticity factor), and equation (8) then reduces to
K = 1.1 ox/Tra/O (9)
It should be noted that the quantity 0.212 results from 11.1)2/(4v'_ T) and that with
consideration of more exact free-surface factors, the value of the expression will change.
Measurements of material plane-strain fracture toughness frequently are made with a
specimen such as that shown in figure 3, the dimensions having the relationships
a Kt/2
2c _ w/3
w _ 6/t
Additional requirements for plane-strain fracture-toughness testing are provided in reference82.
Plots of _2 and Q as a function of a/2c (ref. 83, p. 102) permit easy evaluation of K when
the applied stress and crack dimensions are known. Additional modifications to the
stress-intensity expression that provide for specific geometry effects not covered byequation (2) are covered in references 4 and 83.
The stress-intensity concept has been demonstrated to apply to subcritical (stable) as well as
critical {unstable) crack growth, thus providing a means for correlating the effects of stress
and crack dimensions for cyclic crack growth anti sustained-load crack growth. The
characteristics of critical and subcritical crack growth in several widely used materials have
been extensively investigated. Such data are available in references 84, 85, and 86. The
application of the stress-intensity parameter to tank fracture-control procedure is treated inreferences 2, 3, and 4.
Fracture-mechanics techniques, when properly integrated into a total fracture-control plan,
can provide the desired level of structural reliability. Such a fracture-control plan for a
liquid propulsion system tank includes the following elements:
• Material selection for adequate fracture toughness, in parent lnetal (unwelded)
and welds, under all anticipated conditions of loading and environmentalexposure
• Safe-life analysis based on fracture mechanics
• Quality-control procedures for ensuring both material toughness and detection ofcracks or crack-like flaws
42
i Applicationof fracturemechanicsto qualificationtestingof tanks
• Documentationof all informationandeventspertinentto tankpertormance.
2.2.4.1 MATERIAL FRACTURE TOUGHNESS
A fracture-nlechanics variable important to the initial material selection is the value of the
crack-tip stress-intensity factor K required to initiate unstable crack growth in a particular
material. For a specific material, and for equal conditions of constraint tit the crack tip,
unstable crack growth usually is initiated at a Specific value of K; this critical value Kc has
been designated as the material fracture toughness. For cracks in thick sections, where
maximum constraint (full plane-strain condition) is developed, this parameter has been
designated Kic, the plane-strain fracture toughness. This toughness parameter, which applies
to tension loading normal to the crack surface, is prestlined to be a material constant. When
constraint at the crack tip is lacking, a plane-stress condition exists. In many applications,
the degree of material constraint at the crack tip is somewhere between two cases, and the
crack extension occurs under conditions described as "mixed mode".
The evahiation of materials for tanks on the basis of fracture toughness is based on test data
that represent the specific application. Usually this evaluation requires a series of fracture
tests of precracked specimens with surface flaws (cracks) of varying depths; specimen
thickness is equal to that of the intended application. A high level of material toughness is
desirable: however, the proper material choice involves the relationship of the factors of
material conventional strength, density, and toughness. Figure 4 (adapted from ref. 87)
shows the interaction of these parameters and the region of optimum choices. The "tipper
bound" of figure 4 represents the maximum Ki_ values observed for various material
strength levels, as corrected for material density. The Kl_/oy._ = 0.25 line represents an
attempt to arrive at minimum practical values of fracture toughness, below which the
tolerable flaw sizes may become impractically small from the standpoints of manufacturing
capabilities and the ability to detect the flaws during nondestructive inspection (NDI) as
discussed in the following paragraph. The material choice must also take into account the
effects of anticipated environments on the material resistance to crack growth, and the
crack growth characteristics under cyclic loading conditions; these subjects are discussed
further in succeeding paragraphs.
When a tentative material selection has been made, and material operating stress, thickness,
and toughness determined, the critical crack size can be calculated by the use of the
appropriate equation from reference 88. Normally, the material stress associated with this
calculation is the proof stress rather than the operating stress. (Determination of the
proof-stress level is treated in section 2.2.4.2.) The critical crack size determined in this
manner is the size that is associated with the avoidance of failure during proof testing. The
resulting critical crack size provides a rational basis for assessing the sensitivity of NDI
43
_a c
_c
1000
800
600
400
2OO
o Data points, various structural alloys
\
,, Ti-6AI-4V STA
\ x_ Apparent "upper bound" (current structural materials)
\\o_o_N,,
0
0 0 0 q'_
o o
o o O_o_._ "0000 0 00
0 00.._ _"
0
0
o_
i l I
0
0
0
0
KIc/¢ys = 0.25 -_ °o
200 400 600
• .yield strenqth, ks
density, Ibm/in.3
800 1000 t200
Figure 4. - Material fracture toughness vs yield strength,
density normalized (adapted from ref. 87).
techniques and helps to establish the overall manufacturing and acceptance NDI
requirements. Tile computed critical crack size may be found to be impractically small. Acritical crack size smaller than the sensitivity of the available NDI facilities would increase
the likelihood of tank failure during proof testing. Also, a critical crack size much smaller
than the potential cracks resulting from manufacturing processes could result in excessiverejects during manufacturing and acceptance test. In such cases, the critical crack size is
increased either by choosing an alternate material having a higher ratio of fracture-toughnessto strength or by utilizing a lower operating stress level with the same material.
Another toughness criterion sometimes used in material selection is the "leak-before-burst"
criterion. Usually it is desirable that tank failure be evidenced by leakage without rupture.
This condition requires that cracks that penetrate the tank wall do so without reaching the
critical size for unstable propagation. Tank designs frequently are assumed to possess the
44
leak-before-burstcharacteristicwhenthe calculatedcritical cracksizeexceedsthe materialthicknessby an artibrary factor,which may rangebetweenone and four. tlowever,thereliabilityof thisassumptionisuncertainandvariesbecausematerialsvary in behaviorasthecrack nearscompletepenetration.The plane-stresscondition that norlnally existsasthecrackapproachesa freesurfacepermitsmanymaterialsto undergoconsiderableincreasesinplaslicdeformationprior to separation.With suchmaterials,if the critical cracksizeactisjust equalto or only slightlygreaterthan the materialthicknesst, leakwithout burstmaynot occur. Investigationis requiredto determinethe extraallowancein critical crack size
lamount by which a_r should exceed t) to obtain reliable leak-before-burst performance. Itshould also be noted that the relative advantage of leak before burst behavior is somewhat
lessened when the fhiid contained is hazardous if released.
Fracture-toughness evaluations performed for the purpose of selecting materials usually takeinto account the effects on material toughness of the temperatures anticipated in testing and
service. Materials that are selected are more completely evaluated, consideration being given
to potential effects of all important material variables such as material form, size, directionalcharacteristics, cleanliness, chemical content, thermal processing and deformation
processing. These factors can have considerable effect on fracture behavior. In addition,material characteristics for subcritical crack growth due to sustained loading in fluid
environments and to cyclic loading must be evaluated before a tank safe-life analysis (sec.
2.2.4.2) can be performed.
Environmental Effects
As noted previously, the effects of fluid environment on crack-growth resistance are
determined by loading precracked specimens in the specific fluid environments under
investigation. For most specimen configurations, the initially applied stress-intensity value is
kept below the value that will result in sustained-load growth in an inert referenceenvironment. The highest stress intensity for which there is no crack growth during a
sufficiently long loading period is termed the "material threshold stress intensity" for the
particular medium and is designated KWh. This value of the stress intensity is used todetermine the initial crack size that will grow (and therefore potentially cause failure) in a
particular fluid at a given level of stress. An alternate and more rapid approach fordetermining the threshold stress intensity for crack growth in a particular medium makesuse of a wedge-force-loaded specimen in which the stress intensity decreases with increasing"
crack length. In this approach, a single specimen indicates the stress intensity at which arrest
of growth takes place. Tests comparing the threshold stress intensity for crack-growthinitiation in conventionally loaded specimens with that for crack arrest in
wedge-force-loaded specimens of the same material and exposure medium have resulted in
good agreement. Comparative data of that type are reported in reference 89. A collection ofdata on threshold stress intensities for crack growth in several tankage materials when
exposed to various fluids, as compared with their respective nominal Ktc values, is provided
in reference 4 (p. 21).
45
Cyclic Crack Growth
The application of fracture-nlechanics technology to the problenl of crack growth under
cyclic or fatigue loading has provided useful generalized analytical relationships by which
tile effects of loading spectra on the growth of cracks in materials may be estimated. One
such relationship, from reference 90, is
da/dN = C (AK)" (10)
where
da/dN = crack growth rate (inch per load cycle)
AK = difference between maximum and minimum stress intensity during cyclic loading(Kma x -- Kmin)
C = a material constant
n = a material constant
Another frequently used equation, from reference 91, is
da/dN = C (AK)_/[(I - R)K_ -- AK] (il)
where
Kc = critical stress intensity
R = ratio of minimum to maximum stress intensity during cyclic loading
Either of these equations may be used to predict the growth of a crack from some initial
dimension ai during cycles of tank pressurization, if the material constants C and n
have been properly evaluated. Evaluation of these constants requires a program of cyclic
loading tests on precracked specimens that represent the tank as to material, material
thickness, and fluid environment. Equation (10t differs from equation ill) by the
requirement that these constants must also be evaluated for each distinct loading ratio R.
Equation _ 11 L on the other hand, attempts to provide for the effects of loading ratio.
A further consideration in the use of either of the above equations is the fact that crack
growth characteristics frequently are observed to change as the crack progresses. When this
condition occurs, the material constants must be evaluated for each distinct growth region.
Cases in which three distinct growth regions were identified arc treated in reference 92.
Another method for correlating data on cyclic crack growth, perhaps simpler than those
discussed ahovc, makes use of the ratio of the initial stress intensity (based on the initial
46
crackdepthandtile nlaxinmmcyclic load) to thecriticalstressintensityfor unstablecrackpropagation.The higherthis ratio, the lower thenumberof cyclesof loadingto fracture.This approachhasprovenusefulfor correlatingcrack-growthdatain thecycliclife regionoflessthan 1000 cycles,whenthe critical crackdepth is lessthan the materialthickness.l)etailsof thisapproachanditsapplicationto pressurevesselsarecoveredin reference4.
The wiriousapproachesthat havebeendevelopedfor correlatingcyclic crackgrowthdataundercomplexloadingconditionstend to becomeinaccuratewith increasingloadvariationand increasingcomplexity of loading spectra.Such limitations are evaluated,andrepresentativespectralloadingtestdataareobtainedwheneverit appearsthat theavailablemethodsfor datacorrelationmaynot besufficientlyaccurate.
Regardlessof the method used to evaluate cyclic crack growth, it should be noted that when
loading cycles occur in the presence of a fluid environment other than dry air, it usually is
necessary to consider tile possible effects of such environments on crack growth. Further
information on cyclic crack growth, including considerable test data on materials of interest
for tankage, is provided in reference 84.
2.2.4.2 SAFE-LIFE ANALYSIS
Tile fracture-mechanics treatment of unstable crack propagation, environmental effects, and
cyclic crack growth can be combined to verify analytically the ability of a structural system
to withstand service loading and exposure conditions in the presence of a starting crack of a
given size _rel's. 2, 3, and 4). In such calculations, the initial crack sometimes is arbitrarily
assumed, but a more realistic starting point is to consider the capability of applied
nondestructive inspection. In this approach, the largest crack that could escape detection is
assumed as the worst-case starting crack. A third method for establishing the starting crack
size is to compute it from the stress associated with proof pressurization and the material
stress intensity for unstable crack growth (usually the K_ value). This approach involves tile
assumption that cracks just under the critical size for unstable propagation during proof
testing actually exist in tanks that pass proof testing.
With the maximum possible initial crack size thus established, tile crack growth during
pressurization cycles can be determined as described previously. The reliability of a tank is
verified when tile final computed crack size, at the end of the cyclic life preceding final
service pressurization, is smaller than the size that will grow in the service environment at
tile maximum operating pressure.
By reversing tile above procedure, it is possible to begin with the largest crack that will not
fail in service and work back to the crack size that should just pass proof testing without
failure: from this crack size, together with the known material stress intensity for unstable
crack growth, the required proof pressure can be computed. Further, to avoid failures
47
during proof testing, the computedflaw sizeassociatedwith proof testingbecomesthestandardby whichtlawsthat mustbeeliminatedduringmanufacturingandinspectionaredetermined,sincellaws larger than this size will cause fracture during proof testing. This
procedure for determining the reliability of a pressure vessel or, alternately, for establishing
proof-testing requirements that guarantee tile safe life, termed tile "proof test logic", is
covered comprehensively in references 3 aFid 4.
File ability of a safe-life analysis to prevent service failures may be jeopardized if subcritical
cracks are capable of signil'icant growth without failure during the proof-testing cycle.
Current evidence indicates that there may be many materials, especially those at medium
levels of strength and relatively high fracture toughness, that are capable of such growth
without unstable propagation during a single cycle of loading. With such materials, loading
to the proof stress can produce crack growth without fracture. After unloading and then
loading a second time to a stress significantly below the proof stress, growth can again
occur, this time possibly to a critical size, whereupon fracture occurs. Investigations of this
phenomenon have demonstrated its existence in 2014-T6 aluminum, 2014 alun3inum welds,Ti4_A1-4V STA, Ti-6AI-4V annealed, and D6Ac alloy steel heat treated to 205 ksi yield
strength: the results of these studies are reported in references 03 and 94. With these
materials, or any material suspected of behaving similarly, the conditions under which such
subcritical crack growth occurs and the magnitude of tile growth must be determined and
accounted for in the performance of tile safe-life analysis (tel'. 3, sec. 5).
2.2.4.3 ADDITIONAL ELEMENTS IN A FRACTURE-CONTROL PROGRAM
Quality Control. The toughness of production tank material is subject to considerable
wiriation as a result of tile effects of a large number of material and processing variables.
Only by providing quality control of material toughness can one obtain production parts in
which the toughness values meet or exceed those used in fracture-mechanics analyses.
Quality is controlled by perl'or111ing fracture-toughness tests on unwelded and welded
material obtained from selected areas on production or preproduction parts. In addition, all
material scheduled for the manufacture of tanks is subject to testing to ensure that
toughness requirements are met after production processing.
The elimination of tlaws large enough to affect tank performance, both in incoming material
and in fabricated tankage, is an important factor in a fracture-control plan, because such
elimination prevents or minimizes proof-testing failures and contributes to reliability in
service. Tile most commonly used flaw-detection methods for tankage are dye-penetrant
(for flaws that reach an accessible surface), radiography, ultrasonic inspection, and, for
ferromagnetic materials, magnetic-particle inspection. A review of tile capabilities of these
llaw-detection techniques (as applied to solid rocket motor cases) is provided in reference
16 (pp. 9, 10, 41, and 42).
48
Qualification Testin_g. tlardware qualification programs, performed to verit'y the adequacy
el" design and of fabrication processes, are an important adjunct to a fracture-control
program. Typically, tank qualification 13rograms involve completc physical, chen3ical, and
metallurgical evalttalioi'l of structural components, nolldestructive and dimensional
evahtation of structural test items, functional and structural testing of tanks, and
fracture-meclmnics-bascd analyses of tank fractures. Structural testing of tanks usually
includes cyclic pressurization, sustained pressurization (often while the tank contains the
t]uid to be contained in service), and burst testing. The qualit'ication testing pcrt'ormed to
verify tanks for the Apollo Command and Sen'ice Modules is delineated in refercnce 95.
The performance of actual tankage when cracks or crack-like flaws of various sizes are
present may be evaluated by performing a series of burst tests on preflawed tanks that are
similar in material, geometry, and fabrication processes to the final hardware, but may be
subscale in order to effect testing econolniCS. Tests of this type provide an accurate
reflection of tank fracture toughness by taking into account not only the manufacturing
variables, but also the effects of biaxial loading, curvature, and inaterial thicknesses (which
frequently arc less than the thickness required for development of plane-strain conditions).
Documentation. Maintenance of a system of documentation, sometimes referred to as the
"tank pedigree", provides information necessary for predicting the structural capability of a
given tank at any given time. Complete documentation enables a component or material to
be traced back through all major test, fabrication, repair, and processing steps and associated
inspection records to the original material acceptance test results. It also provides a
continuing record of tank use that contains all pressurizations and conditions experienced
during any operation, including peak pressures, environnlents, temperatures, and number of
depressurization and represstirization cycles: and notes also any mishaps suffered and
corrections made and any other events that may affect the tank structural performance.
2.3 TANK STRUCTURAL DESIGN
Vehicle tanks and the slnaller internally mounted subsysteln tanks have many design
problems in common. They each have weld joints, access openings, support fittings, and
accessory attachment provisions. Both may be required to withstand external collapsing
pressures during preparation for propellant servicing or during decontamination. Vacuum
frequently is used to evacuate gas from the liquid side of expulsion devices and to promote
removal of residual propellants from tanks. To preclude accidental collapse of tank shells,
the rigidity necessary to withstand one-atmosphere external presstire often is incoI'porated
in the membrane design. The massive size and the compressive loading requirements usually
imposed on vehicle tanks, however, introduce design considerations that are not particulcirty
pertinent to subsystem tank design.
49
Minimumweight is alwaysof prime importancein designbut it is by no meansthesolearbiterin ot_timunlvehicledesign.Manytradestudiesaremadeto evaluateweightrelativeto a varietyof otherfactorssuchascost,schedule,producibility,reliability,andawiilabilityof material,technicalskills, tooling, and testingfacilities.Thesefactorsareparticularlysignificantbecauseof the sizeof the final product.To minimizewehling,for example,thelargestpossiblesectionsof raw stock are tised" this practice,in turn, dictatesmassiveequipmentandfacilitiesfor machining,handling,andwelding.
In the structural-design phase, the complete detail design of the vehicle tank including
sidewall, fore and aft bulkheads, access openings, and accessory mounting provisions is
established on the basis of the tank material selected during the configuration-optimization
phase. The major structural junctures (e.g., the skirt/bulkhead/sidewall juncture (fore and
aft), and intertank bulkhead {if used)-to-tank sidewall juncture) are also established.
The design of smaller, monocoque subsystem tanks, although less complicated than the
design of vehicle tanks, is a critical engineering activity. Failure of a subsystem tank under
working stress usually propagates into catastrophic damage to adjacent structure and
components. The continuing necessity for weight reduction in space vehicles has forced
development of minimum-weight, low-margin tanks.
Design of a high-performance tank, where minimum weight is the overriding consideration,
requires detailed structural analysis following the establishment of shape and size (volume)
during configuration optimization. Unlike vehicle tankage, for which material is selected
during the configuration determination phase, the material for the smaller subsystem tanks
usually is selected in the design phase. Weld-joint design, tank membrane thickness, ports,
access openings, tank membrane penetrations, internal accessory provisions, and tank
support provisions also are established. Frequently, initial limited knowledge of tank usage
requirements makes it neccessary to restrict the preliminary design to forging, tooling, and
interface provisions. As information on tank usage becomes more definitive, the final design
analysis is performed.
2.3.1 Safety Factor
The design of a tank usually is defined on the basis of the relation between the loading
conditions that will be imposed on the tank and the capability of the tank to withstand
these loads. Limit load, design safety factor, design load, allowable load, and margin of
safety are tank design terms that are used to define the relation between tank loading and
tank loading capacity. These terms are defined as follows:
Limit load (or pressure): Maximum expected load (or pressure) that will be experienced byt-l_-etank structure under the specified conditions of operation, with allowance for statistical
variation.
5O
Desig21s_ffetyfactor: An appropriate arbitrary multiplier greater than t applied in design to
account for design contingencies such as slight variations in material properties, fabrication
quality, load inagnitude, and load distributions within the tank structure.
De.sign loa_l (or stress): Product of tile limit load (or pressure) and the design safety factor.
Allowable load (or stress): Load (or stress) that, if exceeded, produces tank failure. Failure
may be defined as buckling, yield, or ultimate, whichever condition prevents the tank from
performing its function.
M_argin of safety (MS):
load (or stress):
Fraction by which the allowable load (or stress) exceeds tile design
1MS=-- - 1
R
where R is the ratio of tile design load (or stress) to tile allowable load (or stress).
In general, the magnitude of a safety factor is a rel]ection of the degree of confidence in
materials properties, production processes, and the validity of the predicted usage
conditions. Ideally, design safefy factors for each component in a liquid rocket would be
established by determining analytically the values that would result in the desired
probability of success in the intended application. This safety factor therefore would
incorporate tile effects of variations in material properties, fabrication quality, load
redundancy, and precision of analytical techniques. The design safety factor should relate
mathematically to the desired reliability with an associated confidence level. Theoretically,
each component would have its own safety factor, which would be cost and weight
optimized against all other components to achieve tile desired overall vehicle structural
reliability.
Unforttmately, tile state of the art has not developed to the point where a rigorous
mathematical approach is possible. Instead, a uniform design safety factor for tile entire
vehicle structure is established, and, largely on the basis of experience and judgment, a wide
range of margins of safety greater than zero is used for the various components. Values for
design safety factors in current use for vehicle tanks range from 1.0 to 1.1 for yield and
from 1.25 to 1.5 for ultimate, the higher values being used for manned flight vehicles. In the
smaller subsystem tanks, a wider range of safety factors has been used, probably because tile
tank weight does not increase significantly with somewhat larger margins of safety.
In the design of components such as positive expulsion devices or baffles that perform or
withstand a cyclic type function, the safety factor usually is defined in terms or, and
attained by designing for, expulsion or slosh cycle requirements that exceed mission-related
cycles.
51
2.3.2 Loads Analysis
Vehicle-tank structure must be adequate not only to withstand the loads from vibration,
thermal shock, propellant slosh, and tank internal pressure, but also to provide the main
load path for vehicle body loads. Thus, it is necessary to consider all conditions of loadingduring the design phase, because the type of loads can influence the selection of tank
construction: compressive loads may dictate semi-monocoque construction, whereas purepressure loads may dictate membrane-type construction. Vehicle-tank subassemblies are
subjected to different types of loading conditions and therefore must be examined
separately.
2.3.2.1 TANK SIDEWALL
Vehicle-tank sidewalls are subjected to pressure, propellant inertial forces, axial load, and
bending moments. Loads in the hoop direction are determined by combining the tank ullage
head pressures and load pressure with the propellant inertia forces. The longitudinal loads
result from a combination of ullage pressure, tank axial load, and bending moment. Methodsfor combining these loads are shown in reference 96.
2.3.2.2 END CLOSURE
Loads imposed on the end closures of vehicle tanks are the result of the tank ullage pressure
and acceleration forces on the propellants. During boost, aft bulkheads have a maximum
pressure at the apex, whereas forward bulkheads have minimum pressure at the apex. To
establish the load at a specific location on the end closure, both pressure and closure
geometry must be considered. Methods for determining these loads are described inreference 97.
2.3.2.3 INTERTANK BULKHEAD
Where there are separate and individual bulkheads on two adjoining tanks, the load on each
is determined as described above for tank end closures. The single "common bulkhead",
however, is subjected to either burst or collapse loads and to temperature gradients through
its thickness. Applied loads (pressure) at any given point are calculated in the manner
described for tank end closures, except that the bulkhead "feels" only the difference
between the forward and aft pressures at any given point. To maintain stability under
collapsing pressure loads, the bulkhead is designed to have a large bending stiffness (as
compared with a membrane bulkhead). Consequently, the resulting design (stiffened skin or
sandwich) requires a more sophisticated internal-loads analysis. This generally is
52
accomplishedby amultilayeredshell-of-revolutionprogram(refs.08and99) that takesintoaccountthe extensional,shear,and bendingstiffnessesin both the hoop and meridionaldirectionsof both facingsheetsandcore.A membranebulkheadissatisfactoryit"cautionisexercisedin maintainingapositiveAp in the right direction.
2.3.2.4 ATTACHMENT
Loads are imposed on a tank at the points where other portions of the system arc mated to
the tank. The magnitude and direction of the loads generally are a function of the wcight of
the attached component or subsystem multiplied by the established amplication factor for
acceleration and vibration forces. Only the attachment bolts and the most immediate
structure feel these loads in full magnitude. To a great extent, experience and judgment are
used to determine the magnitude of load on the adjoining structure by making due
allowance for the damping that occurs as the force progresses from the point of excitation.
Axial growth of vehicle tanks poses no restraint on tank function, but attachment of
subsystem tanks must be analyzed carefully to allow for the loads generated at the
attachment points by changing pressure and temperature.
2.3.3 Membrane Thickness
Except in rare cases, liquid rocket propulsion system tanks are thin-wall structures, i.e., the
wall thickness is less than about one-tenth the tank radius (ref. 100, p. 293). This permits
the use of simple stress formulas in the trade studies of simple geometrical shapes such as
spheres and cylinders. Tentative material selections and corresponding membrane
thicknesses and operational stress levels are obtained by using only mechanical-property
data in conjunction with arbitrarily selected factors of safety. Practical considerations of
producibility, handling, and stiffness are evaluated to validate these preliminary selections.
Before material is selected, working stress level, membrane thickness, and fracture-control
consideration must be taken into account. The material must possess suitable levels of
toughness and resistance to subcritical flaw growth to ensure compliance with intended tank
service life. Material selection is followed by definition of the proof-test stress, operational
stress, and NDI requirements consistent with mission performance requirements, as
discussed in section 2.2.4.2. In the determination of membrane thickness, the most
restrictive condition (e.g., safety factor, fracture-control criteria, producibility, or handling)
must be identified and employed.
Because of its strength/density ratio, a certain material may provide a weight advantage
analytically that is negated by further investigation. In trade studies made for the
propulsion-system tanks for the Apollo Service Module, it was determined that fiberglass
53
provided the lightest weight tank. However, the addition of reinforcements for reaction of
the support loads, tank membrane penetrations, and attachment of closeout doors negated
the weight advantage, and 6AI-4V titanium alloy ultimately was used. Sometimes the
analytically permissible thinness of the tank membrane may be outside the limits of proven
fabrication techniques, or it may result in extremely fragile tanks. For example, the shells
for the positive expulsion tanks in the Apollo Service Module required a hemispherical
membrane thickness t6AI-4V titanium) of 0.011 in. based on pressure considerations only,but manufacturing and handling considerations dictated the 0.023-in. thickness that
presently is used.
In the case of vehicle tanks, for which material is selected during configurationoptimization, the designer proceeds directly to establish membrane thickness. The
membrane thickness for a vehicle-tank sidewall is dictated by the product of hoop tension
load under the maximum anticipated internal pressure times the factor of safety. The
sidewall membrane usually is stiffened against buckling by various means such as stringers,frames, or ribs spaced in a waffle pattern, in some cases (e.g., the Atlas and Centaur
vehicles), additional membrane rigidity is attained through the internal working pressurewithin the tank.
2.3.4 Sidewall
Sidewall design is especially important in the design of large vehicle tanks, which generallynot only contain the vehicle propellants but transmit vehicle body loads as discussed in
section 2.3.2.1. The principal sidewall designs for pressurized vehicle tanks are
skin-stringer-frame, waffle, and monocoque construction. Selection of the optimum type of
structure is dependent on the magnitude of the externally applied body loads and the type
of propellants to be contained. Highly loaded sidewalls generally are designed withskin-stringer-frame construction, whereas lightly loaded sidewalls are waflle construction.
Very lightly loaded sidewalls can be monocoque construction but usually must bepressurized for stability.
2.3.4.1 SKIN-STRINGER-FRAME
Integral stiffening is the form of skin-stringer construction that is best suited for propellanttanks. This design eliminates the thousands of potential leaks associated with the mechanical
attachments between the skin and the stringers and frames of conventional construction.
Two configurations of skin-stringer design shown in figure 5 consist of (a) panels in which
the skin, "blade" stringers, and horizontal ribs are machined as an integral unit, frames being
attached mechanically to the horizontal ribs after the panels are formed; and (b) panels in
which only skin and "T" stringers are machined, and frames are added after forming by
54
" Stringer Membrane _ Stringer 7
\ /
Frame(a) "Blade" stringer (b) "T" stringer
Figure 5. - Two kinds of skin-stringer-frame construction.
mechanical attachment to the inboard flanges of the T stringers. Tile blade design is lighter
in the lower load regime, whereas the T-stringer design is lighter at higher load levels.Membrane thicknesses are based on the hoop tension loads under maximum expected
internal tank pressure. Stringer spacing is dictated by local stability requirements. Stringer
configuration and frame spacing for minimum-weight structure are based on general stability
requirements under critical (varies with temperature) axial compression loads in
combination with internal pressures. The material used must be readily machinable and have
good forming and welding characteristics. Welded joints are highly desirable as a means to
avoid propellant leakage. Since tensile design allowables for a weld are lower than those for
the parent material, compensation is made by an appropriate increase in thickness at theweld joints. It is desirable to design so that the skin does not buckle at the design load. Skin
buckling also impairs the reliability of any external insulation and causes sudden changes inthe flexural stiffiaess of the vehicle stage. The stiffener spacing required to keep the skin
unbuckled at limit load is determined from curved-panel buckling data, due account being
given the stabilizing influence of tank internal working pressure. Once skin thickness and
stringer spacing have been fixed, computer programs (ref. 101) can readily optimize the
stringer and frame configuration to provide the load-carrying capability that meets the
required column buckling and general stability requirements. Figure 6 shows a typicalsidewall construction successfully used on the LH2 tank for the Saturn S-II stage.
2.3.4.2 WAFFLE
A tank membrane stiffened by integral ribs that are arranged in a "waffle" pattern is a form
of construction that not only is efficient in the usual loading range but also possesses many
55
_ Numerical.machine.
milled quarter-section
Stringer
Vertical weld
line --_
Typical cylinder section Insulation _ _
t "_ _ Frame
Circumferentialweld line
Figure 6. - Sidewall construction on LH 2 tank for Saturn S-II
stage.
manufacturing advantages. Waffle constrtlction is utilized in the design of the Saturn S-IV-B
stage. The integral rib stiffeners usually are formed by mechanical or chemical milling of the
waffle pattern in a thick plate; mechanical milling is the more efficient method. Design of
the waffle structure usually is dictated by primary shell stability requirements: therefore, a
low density material is advantageous. The plate material usually is easy to machine, and
fabrication is economical. Because of these two considerations, aluminum-alloy plate is a
prime candidate for the waffle plate material. Figure 7 shows two kinds of waffle
configuration, square and isogrid.
In the widely-used square pattern (fig. 7(a)), the rib spacing is such that the individual skin
panels will buckle under the same load that will cause a general instability or failure of the
entire vehicle tank cylindrical structure. The general instability mode of failure is
56
--Membrane
Integra_ ""
(a) Square waffle pattern
Membrane
i_ _ "'" _ St ri nger
Figure 7. - Two basic waffle configurations.
determined as the critical buckling load of an equivalent isotropic cylinder. The waffle
efficiency theoretically increases with increasing plate thickness and decreasing skin and rib
thickness; however, nfinimum skin thickness is dictated by the hoop tension requirements.
The rib spacing and height are optimized on the basis of this fixed skin thickness. Where the
continuity of stiffening ribs is interrupted, weld lands are sufficiently thick to maintain shell
stability. This construction constitutes a fabrication advantage in that all mechanical and
individual rib splices are eliminated.
Another waffle design that has shown excellent promise is the isogrid configuration, lsogrid
is a pattern of integral stiffening consisting of a gridwork of equilateral triangles, as shown in
figure 7(b). This structural arrangement, currently used on the tankage on the Delta vehicle,
has the following properties:
(I) The bending and extensional stiffnesses are independent of the grid orientation.
(2) No coupling between the grid and the skin due to Poisson-ratio effects exists formaterial with v = 1/3. This condition implies that strength calculations for inside
or outside stiffening are identical.
57
(3) Because of properties I 1 j and t 2), isogrid shell analysis is equivalent to that for
monocoque when the proper equivalent thickness and equivalent modulus of
elasticity are used. Because of the simplicity of the analysis, it is COlnparutivcly
easy to optimi/,e and design isogrid structure.
Analytical wcight comparisons of rectangular waffle grid patterns and isogrid patterns have
shown the isogrid pattern to be lighter. In addition, tests have shown that with isogrid
construction, in comparison with rectangular patterns, appreciably higher stresses can be
developed prior to buckling.
Waffle construction offers several other practical advantages: manufacturing is simple
because of the elimination of internal structure: reinforced areas ira the milling pattern for
cutouts or attachment provisions are easily incorporated: and the fact that the skin will
remain smooth and unbuckled precludes problems of debonding of, and damage to, externalinsulation.
2.3.4.3 MONOCOQUE
In general, pressure-stabilized monocoque construction results in the lowest structural
weight. A material with a high tensile yield strength is necessary. Pressure-stabilized
structures such as the Atlas and Centaur vehicles require extreme care during fabrication and
transportation to preclude handling damage. Prevention of buckling ira a pressure-stabilized
monocoque structure requires that the meridional pressure stress be greater than the
compressive meriodional stress created by external loads.
2.3.5 End Closure
End closures significantly affect cylindrical tankage length, and the basic end-closure
configurations are established during the vehicle configuration optimization phase as
described in section 2.1. Not only is the end-closure shape a consideration, but the closure
between two tanks must be optimized: a selection must be made in favor of two separatebulkheads or one common bulkhead.
Theoretically, there are an infinite number of shapes from which to select. In actual
practice, however, closure shapes usually represent a familiar geometric figure or minor
variation thereof. For example, the Titan and Saturn S-IV tanks are designed with
hemispherical end closures, Saturn S-II end closures are oblate spheroids, and the Atlas,
Saturn IC, and Centaur tanks employ ellipsoidal end closures. Such standard geometric
shapes facilitate weight analysis through computer subroutines as presented in reference 5.
Figure 8 depicts some of the surfaces of revolution that have been analyzed by computersubroutines.
58
Ellipsoid
_ 111ips°id
Sphere
Modified ellipsoid
oroid , c- Sphere
Torispheroid
Cassinian
°
Shallow torispheroid Nontangent ellipsoid
Sphere
Biezeno Cycloid Spiraloid
Toricone Hem isphere
I P I
Modified ogive
Figure 8. - Some surfaces of revolution that have been
analyzed by computer subroutines.
59
Stringer
Frame
Membrane
Common
bulkhead
Waffle
constru
TIG welds
Sandwich
construction
Equatorialweld
Sloshbaffles
Anti-vortexbaffles
Sump Aft bulkhead
Figure 9. - Cutaway drawing of the LOX/LH 2 tanks in the Saturn
S-tl stage.
Figure 9 shows the end closures used in the LOX tank of the Saturn S-II stage. Sandwich
construction is employed to resist buckling where compressive loads exist, whereas a simple
nlembrane construction is adequate for the areas loaded only in tension.
An example of honeycomb sandwich structure is shown in figure 10. For given facing
materials and thicknesses, the core depth and density are dictated by the requirement for
preventing primary instahility and "face wrinkling" of the facing sheets. The varialion of
unit weight of the composite sandwich structure with the ratio of thickness of the two
facings is investigated, and the optimunl arrangement is determined.
6O
Face sheet
Adhesive Expanded core
_mpregnated
scrimcloth__
Face sheet
K;>
Fabricated
sandwich
panel
Figure 10. - Example of honeycomb sandwich structure.
Tlle use of sandwich construction for tile intertank bulkhead of booster tanks offers tile
possibilily of significant weight savings. When tank internal pressures are low and shell
instability is tile critical problem, the facing sheets can be designed to operate at stresses
close to the material compression yield stress. Tile total facing thickness is dependent on
stresses associated with maximum tank pressure. Tile facings may be critical either under
ultimate strength requirements in which the hoop tension stresses due to tank pressure are
of primary importance or under yield requirements in which tile combination of thermal
and pressure stresses is of significance. Reference 102 presents typical methods used in
selecting honeycomb sandwich structure.
2.3.5.1 FORWARD BULKHEAD
This bulkhead generally is convex (external surface), loaded primarily by bursting pressure:
a thin sheet or membrane is the usual design approach. 111 a thin-shell desiml, the
01
head-depth-to-dianleterrelationmustbeanalyzedto determinec()mpressivestressesin theknuckle radii to ensurethat no circumferential buckling pattern occurs. These analyses
generally must be confirmed by tests, since information on buckling of heads due to internalpressure is meager.
To achieve minimum weight, tilt shell thickness is tapered so that the entire bulkhead is
operating at the maximum allowable stress level in the meridional direction, regardless of
shape. For minimum cost, however, a constant-thickness bulkhead is more desirable.
Although a single-piece bulkhead is preferred, material size and current forming methods
such as spin-forming, stretch-forming, hydroforming, or explosive-forming make this
impossible. Generally, therefore, the bulkhead is designed to be fabricated by welding
together a single central "dollar" section to a welded subassembly consisting of a number of
gore segments. This practice avoids tile juncture of the multiple welds where the gore
sections meet at a common point. Hoop compressive stresses usually are avoided, the result
being certain restrictions on bulkhead local radii of curvature, which in turn results inheight-to-radius limitations.
Figure 11 shows a typical weight optimization curve for the Saturn S-II stage generated by
the subroutine of reference 5. The curves show graphically the weight variations of tank
sidewall, skirt, and bulkhead as the configuration of the bulkhead is varied. For the loadingcondition of the Saturn S-II LH2 tank, the hemispherical bulkhead results in the heaviest
structure, because of the necessity for a long skirt length.
2.3.5.2 AFT BULKHEAD
Aft bulkheads differ from forward bulkheads only in that hoop compressive forces may
develop under certain conditions of loading. The region from the equator to the liquid level
always will be in hoop compression during filling (assuming no ullage pressure), and
compressive stresses also may occur during flight as a function of bulkhead local curvaturesand the ratio of ullage pressure to acceleration-induced forces.
Aft bulkheads generally require waffle or sandwich construction or other kind of
circumferentially stiffened structure in the upper portion. Minimum-weight design generally
will dictate a shell of revolution with changing curvature and varying geometry. There are no
reliable techniques for the stability analysis for this type of structure under the varying
biaxial load conditions encountered. It is, therefore, mandatory that the analysis be
conservative and that it be complemented with stringent veril'ication testing. An additional
significant consideration is that aft bulkheads must provide for engine feedlines. A central
location is tile preferable choice for a single engine.
62
Er_
5"
.E
op--
6000
5OOO
4000
3O00
2OO0
I000
0 I I :
0.60
S-II shape
.f
t
!
!
I
0.70
Hemisphere
ellipse
Skirt
iTank
/
a
/ W skirt
,
/ W bulkhead
• = . v v - •
0.80 0.90
h
l
.00
h/a
Figure 11. - Variation of tank weight with variation of end-closure configuration
(LH 2 tanks for S-II stage).
63
2.3.5.3 INTERTANK BULKHEAD
The fluids in a bipropellant vehicle are separated physically (and possibly thermally)either
by two separate membrane bulkheads or by a common bulkhead. Figure 12 shows tile two
bulkhead concepts.
Common
bulkhead
Figure 12. - Sketches of two basic types of intertankbulkheads.
The common bulkhead may be either self-supporting or pressure-stabilized. One feature of
dual separate menlbranes is that space for and access to ttle lluid lines coming from the
bottom of the forward tank is provided. Consideration must be given to the special
problems involved in routing these lines either through or around tile forward bulkhead of
the lower tank.
There are several unique features in the self-supporting type of common bulkhead that make
a determination of bulkhead weight versus bulkhead height somewhat more complex. The
bulkhead must be designed for both bursting and collapsing pressures. The bulkhead
therefore is of waffle or sandwich construction to provide stability under collapsing
pressure. If insulation also is a requirement, sandwich construction usually is tile most
efficient, but consideration then must be given to both thermal and pressure stresses.
Convex-vs-concave bulkhead (forward surface) attitudes are evaluated in connection with
the relative magnitude of the burst-vs-collapse pressures. In general, consideration of the
routing of the forward-tank propellant lines makes the convex (forward surface) attitude
more desirable. Trapped residuals in the forward tank are minimized by tile use of
low-density filler material in tile volume below the propellant outlets.
64
2.3.6 Attachment Junctures
2.3.6.1 WELD JOINTS
Weld-joint design is one of the most critical requirements of sound vehicle tankage structure.
Since weld strength is less than that of the parent material, adequate load capacity must be
achieved by increasing material thickness at the weld joint. Ideally, the thicker weld land
would be made symmetrical about the membrane material. IIowever, the expense of milling
both sides and the desirability of maintaining a smooth exterior surface for aerodynamic
reasons are overriding considerations that lead to the eccentric weld land configuration
usually employed on booster tanks. Effective weld-joint strength can in some cases be
improved by shaving the bead; this practice reduces stress raisers, removes the area most apt
to contain flaws, and improves weld-joint ductility.
Optimum joint configuration can be achieved only by an extensive test program for
determining the proper dimensions for the weld-land geometry for each material. Figure 13
,_j , /sk,o L-- ----t_(a) Abrupt termination of weld land
t
weld
Step or taper{b) Stepped or tapered termination of weLd land
Figure 13. - Weld-joint configurations for vehicle tanks.
displays two weld-land designs for 2014-T6 and 2219-T87, two ahmfinum alloys commonly
used in current vehicle tankage. Weld-land thicknesses range from 2 to 2¼ times the basic
membrane thickness, and weld-land widths range from 1.25 to 2.00 in. The 1.25-in. width is
the minimum required to prevent the heat-alTected zone HAZ and the resultant strength
reduction therein from reaching the fully stressed basic membrane. Weld-land thicknesses in
excess of 2V4 to 2V2 times the basic membrane thickness with abrupt terminations (fig.
13(a)) introduce bending stresses and some reduction in joint strength at the point where
65
the basic membrane meets the weld land. This concentrated strength reduction is avoided by
introducing a stepped or tapered section between tile skin and weld land as shown ill figure
13(b).
In the typical stlbsystem tank, the weld joint for assembling tank sections is potentially the
predominant structural discontinuity. The designer's task is to minimize the structural
discontinuity across this joint. Underdesign results in premature tank failure in the weld
joint. Overdesign results in a ring or span of material that is much more rigid than the tank
membrane; the resultant difference in deflection while the tank is under load causes flexure
in the tank membrane and possibly lower cyclic life. Figure 14 depicts some commonly used
preweld joint preparations that have proven highly satisfactory in tank development
programs to date.
I
[
"J" groove
I
_,'/772_,<N-N-N-__m
i
"U" groove
with backup ring
8urndown
(no filler weld rod)
"V"groove
i
i
"U" groove
Figure 14. - Weld-joint configurations for subsystem tanks.
66
Since weld joints inherently are subject to variations in production, tile experienced designer
strives to minimize the linear length of weld joints ill his tank designs. When possible, tank
cylindrical sections ale made by spin forging, thus eliminating longitudinal welds.
"File probable introduction of distortion and oxidation, plus tile need for elaborate
equipment, preclude heat treatment of subsystem tank assemblies following welding.
Postweld processing usually is limited to tank aging to relieve internal stresses caused by
welding. The designer, as in vehicle-tank design, therefore increases the material thickness in
the HAZ to compensate for the reduction in material strength that occurs as a result of
annealing at and adjacent to the weld. Ideally, the gradations in loss of material strength
would be precisely balanced by increased material thickness through a precisely contoured
joint transition. From the practical standpoint, however, tile additional joint efficiellcy and
weight reduction frequently do not justify the attendant analysis and machine contouring
effort, and a straight-taper joint transition is used.
Use of sophisticated chemicals and high-performance components in liquid rocket systems
has emphasized the necessity for fluid cleanliness and for tanks that arc cleanable. It may be
noted in figure 14 that, except for the joint with a backup ring, the joints will be readily
cleanable. In tanks with small ports, where interior access subsequent to welding is limited,
it is mandatory that the weld-joint design require negligible cleanup following weld.
2.3.6.2 BULKHEAD/SIDEWALL JUNCTURE
A critical item in the design of liquid rocket tanks is the method of joining tile major
structural components (e.g., bulkhead to tank sidewall and skirt to tank sidewall). These
junctures normally occur at a common location and almost always are accomplished with an
appropriate fitting. A widely used fitting is a "Y" ring, so named because of its
cross-sectional appearance. Examples of Y-rings, as used on tile Saturn IC and tile Titan II,
are shown in figure 15. A different method, used on tile Centaur, provides an angle for
mechanical attachment of the skirt (fig. 16).
Tile Y-ring type of joint serves tile following functions:
• Provides a structural load path from bulkhead to tank and from skirt td tank.
• Provides a leak-proof container.
• Provides easy inspectability and weld repair prior to skirt installation.
• Provides for mechanical attachment of skirt.
• When properly designed, creates low hoop stresses in tile difficult meridian welds
(three in Saturn IC tank) that splice the Y-ring segments.
67
S Skirt
Tank
bulkhead
Y-ring fitting
Initial bar size
Tank dome
Skirt
Y-ring fitting _
Saturn 1C
Tank sidewall
Tank sidewall _'_
Titan II
Figure 15. - Examples of Y-ring bulkhead/sidewall junctures.
Forward bulkhead
/--- Skirt attachment
f ring /-i IL o I , J
4 ' I "
Tank sidewall
Figure 16. - Example of bulkhead/sidewall juncture used in
Centaur.
68
A principal disadvantage associated with this type of joint is tile difficulty associated with
making the ring-splice meridian welds. This welding usually is accomplished in one of two
(2)
The initial bar stock is rolled to tank contour, the weld joint is prepared, and the
entire cross section of the bar is welded by an extremely large number of passes.
The desired cross section then is machined from the completed bar ring.
The bar stock is machined to the desired cross section and formed to the required
contour. The circumferential splices then are made by one or two welding passes
along each leg and stem of the "Y".
The completed Y-ring then is mated to the bulkhead and tank sidewall by a circumferential
wekt and is mechalfically spliced to the skirt.
The splice-joint buildup used on the Centaur tanks employs simple flat sheets for stiffening
the membrane in the area of the skirt attachment ring.The ring(T-section) is machined from
a circular forging, thereby eliminating welding of ring segments.
This method has the following advantages:
( 1 ) It avoids expensive forgings, machining, and associated tooling.
(2) The bolting ring works with a variety of configurations, and mating of adjacent
structure is comparatively simple.
(3) Weld discrepancies are easily repaired.
It has the following disadvantages:
(1) The method requires spot welding through as many as five layers of skin.
(2) The thin sheets require stringent control of weld schedules.
(3) Vehicle separation requires a circumferential shaped charge to fracture the ring.
2.3.6.3 BOSSES AND SUPPORT PROVISIONS
Attachment of system components and tank mounting structure generally is accomplished
by providing local support points or bosses in tank structure. To avoid tank penetration, the
bosses are made deep enough to accommodate mechanical fasteners. When the basic
structure is made from a stock size of sufficient thickness, these bosses are integrally milled
69
with the basic structure. This process creates "hard spots" that remain fiat during forming
and force the adjoining structure to do additional straining during both forming and
subsequent tank tension loading. Additional local reinforcing is provided to minimize this
excess straining by distributing the strain over a larger area. The reduced eccentricity and
increased distance from the basic thickness to the hard-spot thickness minimize bendingstresses at the discontinuity.
When system attachments are made to thin-shell structure such as a membrane bulkhead,
the use of integral bosses requires a much larger initial material thickness and more extensive
milling. Fabrication of attach bosses also can be accomplished by welding a circularmachined ring (containing the boss) to the bulkhead: however, residual stresses from the
welding can cause severe warpage in thin, compound-curvature bulkheads (less severe for flat
or hemispherical shapes). This warpage can be minimized by the use of hard tooling andclose tolerance parts and by shrink-fitting fitup procedures before welding. However,
scrappage rate is high because of the limited repairability of this type of weld.
The nonintegral support provisions involve welding machined fittings (ports, flanges,
support pads) to the tank membrane. This practice requires either providing a ring of thicker
material in the tank membrane to compensate for the strength loss from welding or loweringthe permissible tank operating pressure. In addition to structural analysis of the designs, the
designer must consider both accessibility for tank cleaning following welding and distortion
of the tank membrane due to weld heat. An advantage of nonintegral fittings is that thetank designer has more latitude in the method of fabrication of the tank membranes.
Although the integral fittings complicate membrane forging and machining, experience hasshown that they are superior to welded fittings because of the absence of heat distortion,
strength reduction, weld-induced contamination, and uncertainty about weld integrity that
attend welded joints. To minimize the effect of integral-fitting discontinuities on the
tankshell, however, the tank designer strives to minimize the number of fittings. Frequently,in gas-pressurant tanks, the fluid ports and tank supports are combined into a common boss.
In liquid propellant tanks, the single access opening is attained by use of a closeout cover
that contains both the inlet and outlet lines. When multiple tanks are used in series, the
inlets usually are connected to standpipes within the tanks to preclude reverse migration ofthe fluids and to ensure series feedout.
Internal standpipes and stillwells in liquid-carrying tanks usually require end supports towithstand liquid slosh impact and to alter vibrational response. These internal supports most
connnonly utilize a slip joint to allow for longitudinal expansion and contraction of the
tank but provide support to the standpipe and stillwells against side loads.
7O
2.3.7 Openings and Access Doors
For purposes of this monograph, the term "openings" applies to all types of ports and
access openings in tile tank membrane. Openings and access doors are treated together
because they represent similar design problems in membrane discontinuity. Ideally, the tank
membrane would have no openings or access doors; realistically, of course, these provisions
are required. Tank openings and access doors for propellant tankage generally are located at
or near the apex of the bulkheads. This location has the advantage of being relatively easy to
fabricate and simple to analyze, and it facilitates entry into the tank interior. The primary
design objective is to prevent leakage throughout the entire life cycle. Two methods that
employ integral bolting rings are illustrated in figures 17(a) and (b): the method employed
on the Centaur for the forward door is shown in figure 17(c).
A bearing lip and oversize holes (fig. 17(a)) will reduce joint rotation due to eccentric bolt
shear loading, whereas a large cross-sectional area in the bulkhead boss (fig 17(b)) will
accomplish the same purpose by forcing most of the load to remain in the bolting ring and
Tank
C
Bearing lip 7
I
(a) Door with bearing lip
(b) Bulkhead boss with large cross-sectional area
oor rin
_,_,i,__ Bulkhea d _ Seam weldmembrane
(c) Centaur forward door
Figure 17. -- Three designs for a tank access door.
71
not _,,othrou_h_ the door. I'iither design can incorporate a door of smaller radius of cu_,ature
than that of the adjoi]ling strtlcttlre ill order to ftlrth(:r reduce ]o;.ids in the door, boss, and
across the loin t (dashed-line conl'igu ration).
lixpericncc has also shown that in the design 1"ol- a radial sealing gasket the final gap
reqt|irements between the door flange and the ring flange must be considered, sincc a local
bending moment is induccd as the torque is applied to the assembly bolts. It is desirable to
select bolt spacing ;.nld seal and flange designs that minimize the final bending moment of
the bolt joint.
Tank openings for system installations create local discontint|ities at the attach points. The
subject is discussed in section 2.3.6.3.
2.4 TANK COMPONENTS
2.4.1 Propellant Slosh and Vortex Suppression Devices
Sloshing of propellants can adversely affect vehicle stability and the integrity of the tank
structure. Bat'tles are therefore provided to damp liquid motion. I11 cylindrical tanks, baffles
generally take the form of fiat rings attached to the structural shell as shown in figure 1:
here they serve the double purpose of providing fluid damping as well as shell stability. For
oblate spheroidal tanks, baffles may take the form of a truncated cone either perforated or
open trussed. These cones generally are supported at the equatorial region and provide for
liquid motion inhibition at the surface level only or throughout its height. A clear
understanding of slosh and methods of counteracting it in large tanks may be obtained I1o111the material in reference 103.
Fluid vortexing occurs at engine feed-line outlets: it is most severe at outlets in tank aft
bulkheads, especially at central locations. Antivortex baffles generally take the shape of
radial vanes either attached directly to the bulkhead skins or cantilevered outward from a
central attachment. Figure 1 shows a method used for multiple outlets and figure 0 depicts a
method employed for a centrally located fluid outlet.
2.4.2 Propellant Positioning Devices
A major problem in designing for a space environment is to provide a supply el gas-lrec
propellant to the engine at the start of each firing cycle. !11 the near-zero-"g" operali(mal
environment of orbital spacecraft, gravitational forccs are negligible, and capillary forces
2
determine the propellant location within tile tank. This capillary action, norlnally away
l¥om thc discharge port, may be augmented by significant periods of adverse accelcl'ations
that muy result from aex'odynamic drag, midcourse corrections, or orbit-adjust l]]:.incuvers. II
the prot_ulsion system is to function prot)erly, these forces must be countered by an
expulsion system that will retain propellants at the tank discharge port.
Surface-force lluid-positioning devices have demonstrated the capability to function
sutisfactorily when a combination of these adverse displacement forces does not exceed the
resisting capillary force of the device. Two basic types of surface-tension systems h_lve been
developed to satisfy various spacecraft propellant expulsion requiremenls. As shown in
figure I 8, these basic designs involve partial lluid control or total fluid control.
Liquidoutlel,,_ _/asinlet
(a) Partial fluid control
A
jA- XX sectooAAoue _ __4.
V. f// "Gas,n,ot% j
_Screeni
i
(b) Total fluid control
Figure 18. - Schematics of two basic types of surface-tensiondevices for fluid control.
73
Partial-fluid-control systems retain a sufficient quantity of propellant at the tank outlet to
supply the engine during restart and to sustain flow until bulk propellant settling occurs.
Partial-control systems generally are limited to unidirectional thrust applications, since they
rely oil engine thrust to provide propellant settling at the discharge port.
Total-fluid-control systems provide for communication from the tank discharge port to the
bulk propellant throughout the operational life of the spacecraft. Examples of both partialand total control systems are discussed in detail in reference 104.
The design of a capillary or surface-tension device is rather complex, because it must
consider not only the magnitude and direction of adverse accelerations but the physical
properties of the propellant such as ,4ensity, viscosity, and surface tension, each as a
function of operating temperatures. In addition, vibration environments during operation
{sine and random) affect the fluid capillary stability at the critical frequency of the
surface-tension device, thereby reducing its capability; reference 105 discusses designpractices for vibration environments. Cleanliness and surface condition of a surface-tension
device also greatly influence its performance. Although relatively complex in design and
analysis, surface-tension fluid-positioning devices provide passive simplicity in operation aswell as reusability.
2.4.3 Propellant Expulsion Devices
In liquid rocket system applications where "g" forces imposed on tank fluids are either
nonexistent or random in direction and magnitude, it is necessary to maintain the fluid
continuously at [he exit port of the tank. This is usually done by either containing or
controlling the fluid with an expulsion device within the tank. Positive expulsion devices are
those that provide a solid material interface barrier separating the presstirizing gas from the
propellant. Actuated by pneumatic pressure, these devices either expand or contract to
expel propellant from the tank on demand and, ideally, at constant flow rate and presstlre.
In addition, the positive expulsion device must control the bulk propellant to prevent liquid
vortexing or gas ingestion. As tanks increase in size, the production of expulsion devices
becomes increasingly difficult. Although a metal-diaphragm type of expulsion device has
been designed and developed for a tank approximately 34 in. in inside diameter and over 55
in. long, it is adviseable to use fluid-positioning devices rather than an expulsion device
whenever possible in the larger tanks. Through currently available production capabilities in
welding and forming thin-gage bellows material, in obtaining uniform dispersion of Teflon
sprayed on forming mandrels, and in precision molding of elastomeric bladders, positive
expulsion devices are feasible for almost any application. Table Ill contains a listing of tanks
that employ various types of expulsion devices. Figure 1_ presents schematics of three basic
types of expulsion devices.
74
Guide ring
Gasinlet\ _ -- ] - @
-LPartially '/
expelledposition
(a) Bellows
Wiper seal-__
Gasinlet%
,-'1c-
-_ /Fluid
__ ]._ outlet
_Partiatly expelled position
iston
(b) Piston
Gasinlet,,_ ._ . ..
Bladder/ - z _ ....
Gasullagespace/ Collector tube
(c) Bladder
Figure t9. - Schematics of three kinds of positive expulsiondevices.
75
Selection of the correct expulsion device approach for a particular application entails
consideration of many factors. There are advantages and disadvantages to the various
bellows, diaphragm, and bladder-type expulsion devices. The comparative importance of
these advantages and disadvantages, however, frequently are subjective with the tank
designers involved. The significant design considerations of the more commonly used designs
(bladder and bellows) of expulsion devices are discussed in detail in the following sections.
2.4.3.1 EXPULSION EFFICIENCY
Expulsion efficiency for a fluid expulsion device is its ability, expressed quantitatively, to
expel the liquid from the tank. One hundred percent efficiency simply means there will be
no residual fluid in the tank after the expulsion device has completed its actuation, i.e.,
expelled volumeexpulsion efficiency =
loaded volume
Associated with expulsion efficiency as a measure of the desirability of a given design for anexpulsion device is the concept of volumetric efficiency, which is the ratio of loaded volume
loaded volumeto internal tank volume, i.e., volumetric efficiency =
internal tank volume
Obviously, the total efficiency of a device is the product of the two, i.e., total efficiency =
expelled volume
internal tank volume"
A characteristic of an expulsion device that must be considered in establishing expulsion
efficiency is its resistance to actuation. A device that requires negligible pressure for its
expulsion actuation is desirable because of its negligible effect on system pressure losses, if a
device has an increasing resistance as actuation progresses, the designer must ensure that the
system flow requirements are compatible with the resultant pressure loss.
Propellant that remains in the convolutions of a bellows expulsion device is an example ofresidual nonexpellable propellant. Bellows with a conventional cross-sectional configuration
have a large radius between the bellows elements; this feature is not desirable, since it results
in a larger quantity of residual propellant. The expulsion of propellant on the outside of the
bellows by extension of the bellows is not desirable, because there is more space betweenthe convolutions in the extended position than in the compressed position. Another cause
of residual propellant in a bellows is nfismatch of the shape of the moving head of the
bellows with the end of the tank against which it nests. Expulsion bellows have beendesigned to maximize expulsion efficiency by using a bellows convolution that nests flat
against the adjacent convolution and a moving head that matches the shape of the tank end.It is of interest that in all but one program the fluid to be expelled has been contained insidethe bellows.
76
The piston type of expulsion device, although extremely simple in concept and theoretically
capable of accomplishi13g 100-percent fluid expulsion, manifests various significant design
problems. The cylinder wall must be sufficiently rigid to maintain acceptable dimensional
relationships under working conditions. The piston seal requires a suitably smooth sealing
surface and, if the seal is pressure actuated, excessive frictional drag against the tank wall
must be avoided. The seal itself must be resistant to wear, chemical attack, and damage by
particulate contamination, so that there is minin3un3 possibility of fluid loss through the seal
as piston actuation progresses.
Nonmetallic bladder devices that fold and crumple (usually randomly) as expulsion proceeds
also trap fluids-within the material folds. In addition, the perforated center standpipe, about
which the device collapses, also adds to residual unavailable fluid. An expanding-bladder
device requires provisions such as ribs either in the tank membrane or in the bladder itself to
ensure that no liquid is locked in by the bladder. The rib requirement in turn contributes to
fluid residuals and lower efficiency.
The diaphragm-type devices that nest in the tank shell present the best potential for
extremely high efficiencies although expulsion efficiencies of 98 percent or better are not
uncommon with both metallic bellows and nonmetallic bladder expulsion devices. In
volumetric efficiency (ratio of expellable volume to tank-envelope volume), however, there
is a significant difference. The flexible-material devices able to assume the shape of the tank
are obviously superior in this respect. Furthermore, the convolute height in the bellows
devices contributes to an increase in tank envelope (volume) without a corresponding
increase in expellable volume.
2.4.3.2 MATERIAL
The choice of material for an expulsion device is closely interrelated with usage
environments and device configuration. In addition to the obvious requirement that the
material be fabricated to the design configuration, two significant material characteristics
are perlneability and compatibility.
Common materials used in the metallic bellows and diaphragm designs are 300-series
stainless steels, high-nickel alloys, and 1100 and 3003 aluminum. These materials can
operate satisfactorily over a wide range of temperature, provide a positive sealing barrier
between propellants and pressurant systems, and are compatible with most liquid rocket
chemicals. Metal-foil diaphragms are frequently used in applications where only one
fill-and-expulsion cycle is required. When repetitive cycles are required, the more substantial
bellows or reversible-diaphragm construction is required.
In the case of nonmetallic bladders, materials have been primarily butyl rubber and Teflon,
Teflon being predominant because it is inert to almost all liquid rocket chemicals. Plastics
such as Mylar, Kapton, Dacron cloth, Nomex paper, and other polyester films have been
77
investigated for use at cryogenic temperatures. Obviously, the choice of a bladder material
for use at cryogenic temperatures is based primarily on its flexibility at that temperature.
Most materials become rigid and crack or tear when flexed at temperatures of liquid oxygen
or liquid hydrogen. The choice of a material thickness is also important, since thinner films
are more flexible than thicker films. However, if the film used is too thin, it will not have
enough strength to resist the bladder stresses or it will be ineffective against permeation ofliquid or gas.
Permeation.- Diffusion of pressurization gas and propellant through the membrane of an
expulsion device such as a bladder is of concern because of the possible detrimental effect of
gas bubbles on engine performance and of propellant vapors on upstream system
components. In a bladder or diaphragm assembly, pressurization gas and propellant
counterpermeate an organic barrier material until the equilibrium point is reached.
Equilibrium occurs when the vapor pressure of the propellant is reached in the gas ullagespace (fig. 19(c)); then the propellant partial pressures across the membrane are balanced.
Should the propellant become saturated with the gas pressurant prior to attainment of the
propellant vapor-pressure balance, the pressurant gas will continue to permeate. Theresulting gas bubbles in the propellant compartment of the expulsion device later will bedetrimental to engine performance.
Gas bubbles can be prevented by use of a nonpermeable material such as metal for theexpulsion device or by use of a thin metallic barrier within a laminated nonmetallic material
such as Teflon. When these approaches are not feasible, it is possible to employ measures at
the system level to prevent gas bubbles. One technique is to control the imbalance in
propellant and gas permeation rates by providing ullage w_lume appropriate for the
differential permeation rate; another is to seed the gas ullage space with propellant toaccelerate the vapor-pressure balance (ref. 106).
Compatibility. - Compatibility considerations involve not merely material being degraded
by various liquid rocket chemicals but also the possible catalytic decomposition of the
liquids by the expulsion-device material. The latter is undesirable because gas bubbles formin the propellant and the pressure rises in the propellant tank.
Hydrazine decomposition occurred in the tanks of the Ranger and Mariner/Mars
midcourse-correction propulsion systems. A pressure rise of 1.5 psi per day was noted when
the test temperature was 125 ° F, but this rise diminished to zero when the test temperature
was reduced to 90 ° F. A tank pressure rise also occurred during flight missions (ref. 107). Itwas found subsequently that the butyl rubber compound and the fillers used in the bladders
(e.g., carbon black) in the tanks were the major cause of the problem. A new compound,ethylene-propylene terpolymer Number 10, was developed; it causes a very low rate ofhydrazine decomposition.
Compatibility considerations must also include chemical solvents used in cleaning andflushing or in testing processes. Referee propellants used for acceptance testing may effect
78
physical property changes in organic materials (refs. 108 and 109) that in tile test
environment can result in permanent damage to the expulsion device. Teflon materials have
been intensively investigated because of their wide acceptance as the preferred material for
expulsion bladders. This preference results from its chemical inertness and stability when in
contact with the chemically active liquid rocket fuels. A standard construction for Tellon
bladders has been a thin-fihn laminate of an inner layer (fluid side) of TEE 30 and an outer
layer of FEP 120, both Teflon materials. This construction was used on the bladders for the
Mariner/Mars 1971 mission. Failure of these bladders during flight-acceptance tests, in
which various solvents were used as referee fuids, led to an investigation to determine the
cause. This investigation identified the sensitivity of the standard Teflon laminate to
solvents such as Freon-TF and isopropyl alcohol and the tendency to incur soh, cnt-stress
cracking at strains below six percent when biaxially stressed. The tests also demonstrated
that the standard laminate is susceptible to stress cracking by N 204 and MMtt.
A new Teflon material, designated "codispersionlaminate", was tested concurrently and was
found to be insensitive to solvent stress cracking. This material replaced the standard
laminate material used in construction of the Tefon bladders. The codispersion laminate
consists of 80 percent TFE 30 and 20 percent FEP 9511. Bladder construction consists of
an inner layer of FEP 9511 sandwiched between inner and outer layer of the codispersion
laminate. The investigation and results are reported in reference 110.
2.4.3.3 DESIGN MARGIN
The capability of the expulsion device to perform in excess of known requirements under
usage environments and conditions is called "design margin". For example, cycle life
beyond requirements and ability to withstand full working pressure and dynamic
environments in excess of mission requirements are indications of design margin.
Cycle life.- Most expulsion bladders and diaphragms fold in an uncontrolled random
manner that results in three-corner folds and creases in the bladder material. This behavior
has been satisfactory for elastomeric bladders operating at room temperature but not for
cryogenic-propellant bladders or metal bladders and diaphragms. Mylar and
polyester/Nomex bladders with collapse-control devices completed 25 expulsion cycles
without failure when tested in liquid hydrogen. The same type of bladders in tanks without
collapse-control devices (random folding) failed after ten cycles.
Metal bladders and diaphragms without some type of rolling or folding control usually have
not had reproducible operational characteristics. In rolling diaphragms, an adhesive between
the diaphragm and the tank shell is used to control the "roll and peel" action of the
diaphragm. Convoluting the diaphragm and telescoping it into one end of the tank also has
been a satisfactory method of diaphragm control.
The most severe folding of a bladder occurs within spherical tanks, because of the amount
of material that must eventually collapse ("wad-up") around the perforated center feedout
79
tube. Insertionof the bladdersinto thetankshellsisalsoacriticaloperationbecauseof thepossibilityof damageto tile bladders.Theinstallationholeisusuallyof minmmnldiameter;hence,thebladdermusthefoldedfor insertioninto thetank.
Orientation of tanks in the associated propellant system is also a significant factor.
Cylindrical tanks with the outlet tube on the major axis and oriented horizontally require
collapse of the bladder and evacuation of gas preparatory to filling. This operation is critical,
since occasionally the collapsed bladder material lying along the lower tank surface will
become entrapped by the weight of fluid being introduced. As tank filling continues, the
bladder must successfully pull out the entrapped folds, stretch, or break. During the
qualification test programs on the tanks for the reaction control system on the Apollo
Command Module, it was also noted that the bladders in horizontally positioned cylindrical
tanks usually incurred a twisting action. This twisting apparently was due to the tendency of
bladder material to fall predominantly on one side of the central diffuser tube as the bladder
collapsed. Subsequent fill-and-expulsion cycles can increase the total angle of twist until
adverse bladder damage occurs.
The foregoing undesirable aspects can be avoided by providing flow-through capability
within the expulsion bladder and by installing the tank with its major axis in a vertical
position. The flow-through provision permits expanding (positioning) the bladder in the
tank shell prior to introducing the fluids. It was determined on the Apollo program,
however, that if the major axis of the tank were too long the expanding bladder would
contact the tank wall during prefill positioning and hang up because of friction before it was
completely nested in the upper hemisphere of the cylindrical tank. This behavior means that
when tank filling ensues, folds of bladder could get trapped by the liquid, thereby stretching
the bladder and resulting in possible adverse damage as tank capacity is reached. This
condition occurred on a N204 tank on the Apollo Lunar Excursion Module. The problem
was solved by undersizing the cylindrical section of the bladder by 1.5 to 2 percent.
A TFE/FEP laminate qualified early in bladder programs has been used for most of the
Teflon bladders. However, tests indicate that the FEP layer fails at 1/10 the number of
cycles required for the entire fihn to fail {ref. 111). Evidence of the failure of the FEP layer
prior to the TFP layer has also been seen on tidied bladders. The new codispersion-Teflon
construction previously described has properties superior to those of the TFE/FEP laminate.
Data based on machine crease tests (ref. 112) indicate that the codispersion has about 100times the flex life of the laminate.
Since current space vehicle applications require only one expulsion under mission
conditions, the requirement for multicycle capability as a measure of design margin is
questionable. Tiffs is particularly true for expulsion devices that degrade with each full
expulsion. Therefore, the designer should establish cycle requirements based upon the
expected ground servicing and testing conditions.
8O
Differentialpressure capability. Since expulsion devices effect a positive barrier between
tile fluid and pressurant sections of a system, it follows that full working pressure will be
imposed across the device as liquid depletion occurs. High differential pressure is damaging
to most expulsion devices. Teflon bladders such as those used ill the Apollo spacecraft tanks
fold randomly as expulsion progresses, striate severely and, occasionally, pinhole when a
sustained differential pressure of 185 psi {working pressure) is applied. This problem was
resolved by limiting tile differential to 40 psi during system checkout activities. In general,
resilient materials such as rubber are more resistant to this type of damage.
Tile allowable differential pressure across a metal bellows in its extended position is limited,
since an excessive pressure may cause tile convolutions to buckle. Partial buckling of the
convolutions may markedly reduce bellows cycle life, and complete buckling of the
convolutions may prevent the bellows from recycling (ref. 113). Although the bellows can
take a very high differential pressure in the nested position without failure, complete
compression under high pressure may reduce its cycle life. Therefore, a mechanical stop
usually is incorporated to prevent complete nesting.
Dynamic environment capability.- Vibration, acceleration, and propellant slosh during
launch of a vehicle usually are the most severe that occur during tile mission. Vibration
frequencies range from 30 to 2000 Hz; acceleration is seldom below 5g and may exceed 10g
(ref. 107), with shocks of moderate level (<100g) and duration. At this point in a mission,
the subsystem positive expulsion tanks are fully loaded except for required ullage. The
added mass results in beneficial damping of the vibrational excitation induced in the tanks,
and slosh is minimized by the full-tank condition. It has been determined, however, through
testing with transparent tanks, that flexible bladders incur substantial flexing at tile
liquid-ullage interface during vibrational excitation of filled tanks; it is desirable therefore to
provide resistance to flexure damage.
Resonant frequencies for most positive expulsion devices occur at low to moderate
frequencies (<100 ttz) in the vehicle bending modes. Bladders and diaphragms of organic
materials are particularly susceptible to damage in the low-frequency lateral slosh modes
(ref. 1141. This characteristic becomes increasingly significant in large tanks with moderate
to large ullage volumes (ref. 115). Metallic units can experience impact and abrasive damage
in this frequency range t'ref. 116).
Tile metal-bellows assemblies for propellant tanks on the Minuteman III vehicle were
subjected to extensive vibration and shock tests. The initial 6-mil stainless-steel bellows
assembly provided insufficient spring rate to resist buckling loads imposed by the 30-g,
18-millisecond terminal-peak sawtooth shock test. Also, tile design pitch/span ratio was
excessive, tile result being an operating pitch too close to the critical buckling pitch. The
design changes to solve this failure mechanism included an increase in metal thickness to 7
mils, an increase in the number of convolutions to decrease tile operating pitch, and the
incorporation of a guide pin that limited the rotational moment of the movable bead.
81
Acceleration is of minor concern with flexible-bladder expulsion devices, since they do not
as a rule have to resist the fluid motion. Lateral acceleration of a tank with a bellows device,
however, can produce unsymmetrical fluid pressures on the movable head of the bellows.
This condition can cause the head to tilt at an angle to the axis of the bellows and thus
cause some bellows convolutions to be overextended and to buckle. This problem in the
Mimlteman post-boost control-system tanks was resolved by welding a guide pin to the
center of the movable head that fits into a recess in the end of the tank. The pin, along with
a circumferential guide ring (fig. 19(a)) around the bellows head, prevented the head from
cocking to any extent that would cause overextension of the convolutions.
2.4.4 Tank Insulation
The insulation design for structures used for cyrogenic fluids is of considerable importance
because of the effect of cryogenic temperatures on material properties. Insulation is appliedeither internally, as on the Saturn S-IV-B stage, or externally, as on the Saturn S-II stage.
Internal insulation is used when it is desired to retain room-temperature material properties
for the tank structure and thus avoid the loss of material ductility and toughness that may
occur at very low temperatures. External insulation permits the tank structural design to use
the increased tensile strength of the material that exists at cryogenic temperatures.
Three methods are generally employed for attaching the insulation to the tank walls:
bonding, bolting, or spray foaming. The bonded and spray-foam insulations require careful
surface preparation prior to installation, whereas the bolted insulation requires integral
bosses in the tank structure; design of these bosses is discussed in section 2.3.6.3. Examples
of the installation of internal insulation, bonded; external, spray-foam insulation; and
external, mechanically' attached insulation are shown in figures 20 and 21. An example of
insulation of the Centaur intertank bulkhead is shown in figure 22. Moisture condensation,
formation of ice, and loss of insulation performance or possible insulation damage requirethat extreme care be taken to ensure that external insulation is resistant to moisture or
protected from moisture.
2.5 TANK DESIGN ANALYSIS
Design analysis, the key to a successful tank design, is the analytical prediction of the
behavior of a vehicle or subsystem tank when subjected to both structural and dynamicloadings. This design analysis is performed concurrently with the detail design of tanks. The
general procedure in tank design is to lay out the basic structural concepts and size thestructure from the design analysis.
82
Rub coat sealer
Liner (glassfabric)
Core foam blocks
Tank wall
(a) Internal bonded insulation (S-IV-B)
_'--']A _ Forward skirt
=_ Lower cyti nder
Section B-B Detail A
(b) External sprayAoam insulation (S-II)
Figure 20. - Examples of internal and external insulation.
83
U
E_
L
I
E
I,..- ._c
_ ceO--.__
<,
,<
E
c-
t-
O
E
I
LJ.
84
Monocoque membrlne,
bulkhead"_lnsulati0 n
bulkhead
Figure 22. - Insulation of intertank bulkhead (Centaur).
2.5.1 Strength Analysis
Large vehicle tanks constitute sophisticated structures that are subjected to a multitude of
loads (both magnitude and direction) during use. The smaller subsystem tanks of
monocoque design are subjected primarily to internal pressure and concentrated loads at the
support points: structural analysis for these tanks usually is not as extensive as that forvehicle tanks.
When the vehicle structure is analyzed to demonstrate adequate load capability, all elements
of the structure undergo analysis by cost-effective techniques suitable to each particular
application. Empirical methods developed for analysis of structural assemblies are used
whenever possible, and theoretical methods, supported by special tests, are used when
necessary.
in the usual analysis of shell structure for vehicle and subsystem tanks, elastic behavior forboth limit and ultimate loads is assumed. The analysis is accomplished with the aid of
computer programs for repetitive or complex mathematical procedures. The use of programs
developed to take into account the effects of plasticity in a biaxial stress field is extremelydifficult and rarely is warranted. The uncertainties of material properties in the plastic zone
make the results dubious at best. A simplified linear-elastic theory produces results that are
conservative and not significantly different from those obtained by much more
sophisticated techniques.
85
2.5.1.1 TENSION-LOADED STRUCTURE
The parent material of the basic shell structure is analyzed for tensile yield or failure
according to the basic stress and strain equations as given in any text on strength of
materials (e.g., ref. 11 7). Tensile rupture may result from pure tensile loading or from the
effects of combined bending and axial load. Margins of safety are determined by computing
stress ratios and interaction curves as defined in reference 8. Allowable bending stresses are
based on the "modulus of rupture" as determined by test and are a function of the materialand tile shape of the cross section (ref. 1 18).
It is difficult to perform an accurate stress analysis of eccentric welded joints that combine a
known axial stress with an indeterminate bending stress. The weld nugget yields early, has
no clearly defined proportional limit, and the major portion of the stress-strain curve lies in
the plastic range. For the typical eccentric butt-welded joint shown in figure 23, high local
yielding takes place at location A in the parent material and at location B in the weld, thusrealigning the load through the weld nugget.
j/Weld _B (local yielding in weld)
' " --A (localyieldingin
parent material)
Unloaded joint Loaded joint
Figure 23. - Realignment of an eccentric weld joint under load.
Consistently reliable results have been obtained for large vehicles by basing allowablewelded-joint strength on test coupons of representative configurations. Normal
manufacturing imperfections such as porosity, mismatch, and oxides are included in the test
specimens so that the statistically reduced allowables are realistic. These allowables are
further reduced by 10 to 20 percent to account for biaxiai and curvature effects and
possible unknown scaling effects. The presence and growth of flaws during life cycle
loading, their evaluation, and their relation to proof-pressure loading are discussed in section2.2.4.
86
2.5.1.2 COMPRESSION-LOADED STRUCTURE
The skin-stringer-frame type of construction (sec. 2.3.4.1) requires analysis of the skirl
panels for local btrckling and a determination of effective skirl to be used with tile stringersas columns between frames (ref. 119). A conservative column-end fixity coefficient of 1.0
can be used, or larger values can be established by test for any given configuration (ref. 90).
The inboard caps of stringers are checked for local crippling under a combination of
maximum shell axial compressive load and maxinmm internal pressure. The Poisson effect
of the internal pressure causes compression in the stringers, and the restricted radial
deflection at the frames causes stringer bending. This bending adds maximum compression
to the inboard stringer cap midway between frames. Analytical procedures are given in
reference 120. General stability is accounted for by sizing and spacing the frames according
to methods provided in reference 6.
The waffle type of shell construction (see. 2.3.4.2) subjected to compressive loads and
internal pressure is analyzed for the general stability mode of failure in the same manner as
an equivalent monocoque cylinder. The skin panels are checked for Local buckling, and the
ribs are checked for crippling. Methods of analysis are given in reference 121.
To prevent buckling in a pressure-stabilized monocoque structure (sec. 2.3.4.3)the
meridional pressure stress must be greater than the meridional stress created by external
loads. In the case of the Atlas vehicle, incipient buckling is allowed to occur at limit load.
Thus, at ultimate load, the tank skin is in the buckled condition for a portion of the
circumference. This postbuckling strength is available only when most of the imposed load
is a bending load. Methods for analysis of postbuckling are given ira reference 122.
2.5.1.3 MAJOR JUNCTURES
Major junctures are analyzed by computer programs for axisymmetric loading conditions.
Critical loads caused by the worst combination of bending and axial loads are assumed to
act uniformly arotmd the shell. Figure 24(a) represents a typical Y-ring juncture, and figure
24(b) represents a bulkhead-sidewall juncture for a monocoque structure (of. figs. 15 and
16). The schematic portion of figures show the various elements of the structural assemblies
that are subjected to analysis.
Input data consists of geometry, temperature, and hoop and meridian loads. Elements aredivided into short shells, semi-infinite shells, and rings. The effects of axial loads are
included in the determination of shell deflections and rotations (ref. 123). The method of
analysis is based on matching deflection, rotation, shear, and moment at all element
junctions and is fully described in reference 124. Program output includes discontinuity
shears, moments, deflections, and rotations at all junctures.
87
Weld land
F:x,bla _._w UF..j._av..im,..F walI
LH 2 tank skin
Intertank
bulkhead
[][]
2?LOXtank
skin 72
(a) Y-ring (b) Monocoque
Figure 24. - Structural elements of bulkhead/sidewall junctures
subjected to analysis.
2.5.1.4 LOCAL ATTACHMENTS AND OPENINGS
Axisymmetric openings and attachments in the tank structure usually are analyzed by
considering the opening or attachment reinforcements to be rings or short thick cylinders.In designs involving nonaxisymmetric openings or attachments that result in noncircular
rings, the usual approach is to design a uniform reinforcement for the maximum loads that
would exist on an equivalent symmetrically loaded structure. A finite-element computer
program (ref. 125) has been developed that will facilitate the analysis of the reinforcingpads, nonaxisymmetric openings, and shell-supported rings when reinforcement thicknessesdo not exceed approximately four times the tank wall thickness.
Methods for determining the effects of local loads on tank structure can be found in
references 126 through 129.
2.5.2 Structural Dynamics
Tanks such as large integral vehicle tanks must withstand various aerodynamic and
acoustically induced loads and environments in addition to service temperature and pressure
88
environments. The smaller internally mounted subsystem tanks usually are intentionally
isolated from aerodynamic loads through protective surrounding structure and supports:
complete isolation from the dynamic environment, however, is extremely difficult to
achieve. In tile design of expulsion devices, vibration response, fluid slosh, and acceleration
loads are major considerations.
2.5.2.1 BENDING FREQUENCY
Generally, tile most significant dynamic problem in vehicle design is the proper control of
the body bending frequencies. Bending frequencies influence the aeroelastic behavior of the
system, the dynamic loads resulting from wind gust and steady-wind shear, and the
interaction of tank bending with tile guidance and control system (ref. 130). The
importance of bending frequency depends on the location of the particular stage in tile
vehicle and most often is critical in the center stage.
2.5.2.2 EXTERNAL DYNAMIC ENVIRONMENT
Dynamic bending moments resulting IYom steady-wind shear gusts and buffeting during
flight are influenced by vehicle configuration (e.g., L/D), payload geometry, control system,
and the vehicle flight-profile and operating parameters. Transient dynamic loads usually are
analyzed to determine whether the loads are more or less than the quasi-static loads
determined by rigid-body analysis of gust transients and steady-wind shear and whether tile
dynamic behavior of the vehicle acts as a gust-load alleviator.
External loading conditions that may induce dynamic response in a tank consist of the
acoustical field generated by the engine, aerodynamic or boundary-layer noise, and the loads
encountered during transportation and handling. Acoustical and boundary-layer noise nltlst
be considered in the analysis.
The dynamic behavior of clustered motors is influenced by tile tank stiffness, not only in
the overall vehicle dynamic behavior but also in dynamic interaction between tile motors.
With clustered motors, additional dynamic loading can be generated because of
nonsimultaneous motor ignition and burnout. TVC inputs, and aeroelastic conditions.
Considerable attention has been given in recent years to the dynamic analysis of clustered
structures. At the present time, the analyst has a choice of several procedures (refs. 131
through 134) that include matrix techniques or continuous-mechanics methods.
2.6 TANK FABRICATION
In the design of tanks and particularly tile integral propellant tanks for large vehicles, the
consideration of fabrication procedures and processes must be continually integrated into
89
tile designasit evolves.Present-dayproductionof small-andmedium-sizesubsystemtanksof monocoquedesigngenerallyinvolvescomparativelyroutine forging,machining,andweldingof two or three tank sections. Tile fabrication of a sophisticated vehicle structure,
however, involves fabrication and assembly of many large structural subassemblies. The
major problem is tile size of tile sections that must be processed. For example, a typical
sidewall section of tile Saturn S-II LI]: tank is 103 in. by 325 in. and weighsT0001bm at
tile start of machining. This large size obviously has significant bearing on machining,forming, handling, and assembly processes.
Machining usually is accomplished by tape-controlled, multihead end mills. Forming a large
fiat section into a curved segment of a cylinder is a particularly difficult problem if the
panel has waffle-type stiffening. Integral circumferential ribs used in the waffle construction
create the need to t'orm by a brake forming process. This type of construction is used on the
Saturn S-II stage and, initially, resulted in extensive cracking at the rib ends. Where only
longitudinal stringers are used, the sidewall section may be mechanically attached to a
forming mandrel and heat aged. Tile weld tooling required for joining large-diameter
cylinders is necessarily a major and crucial item it mismatch is to be held to a minimum.
Bulkheads arc formed by either stretch forming or explosive forming; for a thin-naembrane
bulkhead, the stretch-forming technique is the most cost effective. To attain further weight
reduction, the thin-membrane bulkhead in the Centaur LH2 tank is chem-milled, thickened
areas being left to serve as integral ribs and stringers for structural stiffening. Bulkheads that
require stabilization such as that provided by waffle construction generally are formed
explosively. In these cases, the ribs are stabilized with a low-melting-point alloy (e.g.,
Cerrobend) prior to forming. Forming may require two or more stages, with intermediate
annealing. Following forming, the sections arc solution heat treated and quenched for
desired properties, leach of the major fabrication processes presents significant technical
problems, and close coordination between production and design presonnel ensuresoptimum solutions.
2.7 TESTING AND INSPECTION
In tank development programs, testing is used in various phases of tank design, material
evaluation, and evaluation of fabrication and inspection processes to establish design
requirements, to evaluate alternative approaches, and to verify end results. When designs,structural materials, and fabrication and inspection processes are substantiated on the basis
of past experience, testing usually is limited to the degree necessary to certify the new
design. When determination of actual strengths and margins is required, destructive testing is
employed. The sizes and shapes of destructive-test specimens are virtually limitless, and the
specific requirements depend on the particular program needs. The types of destructive tests
most frequently used in tank development programs include hydroburst testing (either
subscale or full-scale), mechanical-property tests with uniaxial and biaxial specimens, andmetallographic and composition analyses.
In addition lo uxlrc'me care i_3 tile design of tanks and accesSoli0s, the Iin:_l product also
requires intensive inspection processes to provide the desired cont'idcn_:e. The l]'lOSt COIlllllOn
inelllotJ of estahli>,hing and maintaiilitlg reliability of :t lank is to employ a _.'ollltwehensive
progl-aln of process control throtlghotit material proct_rement and fabrication. This program
permits detection el" potential causes of failtlrc ail.d timely correction. The degree of st.iccess
of this m¢lllocl, of course, is totally dependerit on the suitability of the iilspection criteria.
One of the I_rimary tests iinl)osed on liquid rocket tanks is tile proof-I_ressure test, which has
ser_'L_'cl for lnany years as one of tile final inspections prior to service usage of tanks. I)uriilg
tile l)tlS{ 1ci1 years, fracture-nlechanics studies _.tlld aerospace pressure-vessel experience have
shown thai a properly designed proof-pressure lest probably is the most reliable
nondestructive inspectioll method available for ensuring reliable htrik service. It is
incunll_ellt on the tank designer to ensure that the lank design and the t_roof-ttrc'stqiro lost
arc conlt_alil_le, i.e., that the test ctonlonstrates adequate service Iif0 of the tank and that no
dalna,ge is incurred by the tank durJn7 test. Procectures for ctesigning an adequate proof lost
tire clescribud in references 3 and 4.
Another reason for perforining adeql_iate arid woll-docunlented inspectioll arid testing of
aerospace tanks is the sot of stringent railge sal'et7 roquirelnents iinposed on liquid rocket
tanks clurillg checkout and launch o1 space vehicles.
91
92
3. DESIGN CRITERIA and
Recommended Practices
3.1 TANK CONFIGURATION
3.1.1 Vehicle-Tank Optimization
Vehicle-tank optimization shall define the minimum-weight tankage structure that
will contain the propellants and transmit the launch vehicle loads.
The largest influence on vehicle-tank weight is the material; therefore, tank material should
be selected early in the optimization activity. Various tank end closures then should be
analyzed to identify the configuration that provides the required fluid capacity withminimum tank weight. Finally, preliminary sidewall designs should be compared in terms of
ease of fabrication, past usage history, material size availability, and weight. The variouscombinations of materials and general bulkhead and sidewall configurations should be
compared to identify weight trends and to identify the optimum configuration with greater
certainty.
3.1.2 Subsystem-Tank Optimization
Subsystem-tank optimization shall define the minimum-weight tank .for the
prescribed usage and envelope.
The optimization procedures necessarily involve system configurations. The recommended
procedures for the designer are as follows:
(l) Convert the basic rocket-system requirements (e.g., specific impulse) into fluid
quantities.
(2) Identify the increases in fluid quantities that must be included to compensate for
system-created "losses". In the case of gas pressurant, the resultant loss in volume
due to temperature reduction during gas flow must be determined. Accuratesolutions of the heat-transfer and energy equations for the majority of stored-gas
systems that pressurize (downstream) in a transient-temperature manner are bestmade in incremental steps during graphical or computer programs. For
preliminary design purposes, use much quicker approximation methods for
93
(3)
estimating gas temperature within an associated liquid tank (ref. 135). Consider
fluid leakage, ullage space (liquid tanks), expulsion efficiency, filling errors,
feedout imbalance, and manufacturing tolerance ill the inventory of incremental
volumes. Sum these increments and add to the basic tank volumes to obtain totalminimum volu me.
Use the above-determined tank volume and any imposed installation constraints
to compare various geometrical shapes: select the best configuration. Comparison
considerations should include the associated tank-supporting bracketry.
3.2 TANK MATERIAL
3.2.1 Mechanical Properties
The values )br _tulterial mechanical properties used in tank design and analysis,
together with the design ./actor of saJet),, shall provide an adequate and consistent
level of reliabiEo, against nulterial yielding and ductile failure.
All design values of material mechanical properties Ftu, Fty, Fcy, F_u, Fbru, andFbry should be established at consistent levels of reliability that include consideration of
both the scatter in material properties and the eft'ects of all processing, environmental, and
service variables. Unless special conditions dictate otherwise, the reliability levels used
should be similar to the "A" and "B" levels in reference 8. "A" values should be used in the
structural analysis of pressure vessels, because these vessels represent single-load-path
structures, the failure of which would be catastrophic.
Material design data used in fatigue and creep analyses should be established by methods
that take into account both the variation in material properties and the analytical approachto be used.
3.2.1.1 TEMPERATURE EFFECTS ON PROPERTIES
Tank material properties s/tall be adequate .fbr operational and test kmdingconditions under all aJttieipated thermal em, ironmen is.
The material properties used for the design and analysis of tanks that operate at elevated
temperatures must take into account both the temporary effects of increased temperature in
reducing properties and, when temperatures are sufficiently high, the permanent strengthreductions due to thermal exposure.
94
The materialpropertiesusedfor designand analysisof tanksthat normallyoperateatcryogenictemperaturesmustbeadequatefor all specifiedandpredictedroon>temperatureoperationalor test conditions. For many tank applications, it is impnictical to take
advantage of the higher mechanical properties at cryogenic temperatures, because of the
restllting necessity of placing an extra limitation on tank pressures for roonl-temperature
testing or operations and, sometimes, also because of an undesired reductio11 in the ratio
between fracture toughness and material design strength at cryogenic temperatures.
3.2.1.2 FATIGUE STRENGTH
Tank mater&Is shall withstand the specified multiple load c roles.
Tank design, material selection, analysis, and development test programs must take into
account the maximum number of pressurization or other load cycles to which the tank
might be subjected. Besides service conditions, these cyclic loads should inchide all
proof-pressurization, leak, and operation testing of tank and system both before and after
delivery of completed systems. Materials should have adequate low-cycle notched fatigue
strength both in parent metal and welds. Stress and strain concentration should be
nfinimized in design and manufacturing, especially in weld areas. Vessel development and
qualification test programs should require demonstration of ability to withstand the
maximunl number of presstlrc cycles that will be imposed during service life.
3.2.1.3 CREEP
Ta_zk materials shall not dejbrm or rupture as a result of creel).
The possibility of material creep should be considered in all tanks that are pressurized at
temperatures above room temperature. Because of the titanium low-temperature creep
phenomenon, a creep analysis should be performed on all titanium-alloy tanks. The
possibility of creep during testing procedures involving titanium tanks should be considered.
Creep should also be checked for aluminun_ tanks that are pressurized at room temperature
for a total duration exceeding 1000 hours.
3.2.1.4 BIAXlAL-STRESS PROPERTIES
Tan k material shall withstand the an ticipa ted biaxial loading.
Since the various materials of interest for tank construction do not necessarily behave in
accordance with any single, idealized theory of material deformation or failure, it is
necessary to evaluate biaxial properties from representative test data. In performing such
evaluations, the following factors should be considered:
95
(1)
(2)
(3)
(4)
The stress state used to determine possible biaxial effects in a tank must be the
actual stress state at the failure initiation site, including the effects of bending
stresses and stress concentrations, if such exist.
High stress concentrations that are critical failure-initiation points can locally
obliterate the general biaxial stress field together with its potential effects.
Tile evaluation of biaxial effects must be concerned with the element or portion
of the tank that is actually critical, whether a weld, or weld heat-affected zone.
the area around a boss, or any other similar portion.
The biaxial effects observed for initial yielding should not be expected to be tile
same as those observed for ultimate failure, and vice versa.
3.2.2 Fabrication Considerations
Tank materials shah admit of being fabricated into the necessary configurations
and sizes within the limitations of cost and schedule, when fabricated, the
materials shall possess the physical properties used in design and analysis.
Manufacturing feasibility studies should be performed to aid in the selection of materials-
and manufacturing processes that will make possible the production of tanks having the
desired configuration, properties, and quality within the limitations of schedule and budget.
Such studies should take into account the capacities and capabilities of available equipment
for performance of metal forming, machining, welding, heat treating, and other essential
processes as compared with requirements for such as indicated by the proposed design,
method of fabrication, and material characteristics. The cost and schedule impacts
associated with the construction or provision of such facilities that are otherwise unavailableshould be considered.
3.2.2.1 SHAPING AND FORMING
Tank n_terial s/tall, on completion of all de fbrmation processing and other
fabrication processing, exhibit acceptable qualiO' and mechanical properties.
The quality and properties of material that has been shaped, formed, or otherwise fabricated
into tank components should be verified by a complete testing program for qualification of
components. The material chemical composition, structure, soundness, grain size and flow,
freedom from defects, and mechanical properties should be verified throughout such
components. Provision should be made to provide continuous surveillance of the quality of
96
such components through inspection and testing material provided for this purpose, either
integral with the components (e.g., forging trim rings) or, if such cannot be provided,
material processed along with components.
3.2.2.2 WELDING
Welded ]oints in a tank shall exhibit consistent and adequate levels of strength,
ductility, and toughness.
Alloys selected for fusion-welded tanks should be suitably weldable and weld reparableunder all welding conditions anticipated. Welded joints must have adequate strength,
ductility, and toughness at all temperatures at which the tanks are to be pressurized.Materials for which weld joints are considered to have limited ductility should not be used
in designs that incorporate (1) longitudinally welded cylindrical sections, (2) domes weldedfrom segments, or (3) welds located in regions of stress or strain concentration (e.g.,
welded-in bosses and attachments).
To verify that satisfactory levels of joint soundness, freedom from defects, strength, and
ductility can be met consistently, weld development studies should be performed to
optimize welding conditions for the specific joint materials and geometries contemplated.
Weld quality-control procedures must be established to ensure the maintenance of weld
property levels and freedom from defects.
3.2.2.3 THERMAL PROCESSING
The material thermal processing requirements shall be satisfiable, within the
limitations of budget and schedule, jbr the sizes and configurations ofactual parts
or completed tanks. At completion of all thermal processing, tank materials shall
exhibit the property levels used in design and analysis.
In selecting materials for tanks, it is necessary to verify that any detailed thermal processing
requirements can be met for the sizes and configurations of the actual parts or completed
tanks for which such processing is contemplated. The need for such processing may arisefrom the need for performing various manufacturing operations at low material strength
levels or from the need for thermal processing after welding. It is necessary to determine
whether the indicated thermal processing can be performed by existing facilities or will
require the construction of special facilities.
The material properties and microstructure obtained from heat-treat procedures andfacilities should be determined to be satisfactory by thorough destructive and
nondestructive testing of material from initial production hardware. Quality-control
97
procedures should be established to maintain all important heat-treatment parameters
within the necessary tolerances. Such procedures should include the testing of representative
coupons processed together with the parts or completed tanks.
3.2.3 Material Compatibility with Environments
Tank materials shall m_t he degraded uHaccelmzhly, become embrittled, or .l_til
i_remutttrelv as u re,suit of interaction with aggressive .[htids, I_roces,s'e,s, _)r
environmeHts: and l_ropcllaHts shall not he co#;tumi;mled or decomposed by
reaction with tank materials.
Alloys must be evaluated lk)r compatibility with the fhlids to be contained and external
environments to be encountered in service; in addition, evaluation must include all other
chemically active or otherwise suspect substances to which materials are exposed during
manL_facture, storage, testing, and transportation. Substances and processes that are found
to be deleterious should not be used on tank materials.
To evaluate proposed alloys and the fluids, processes, or environments to which the alloys
arc to be subjected, data on the following possible modes of material or system
deterioration or adverse reaction may be required:
• Corrosive attack from manufacttLring fluids and processes, testing fluids, fluids
stored, and atmospheric or other external service environments
• t{mbrittlen3ent restllting fronl internal contamination introduced by
lllantifacturing fluids and processes
• Formation of undesirable products of metal/fluid reactions
• Promotiol, of propellant decomposition (metalftquid reaction)
• Stress-corrosion cracking
• Galvanic corrosion
• Hydrogen-environment embrittlement
• Material ignition in propellants.
98
3.2.3.1 SOURCES OF MATERIAL/FLUID REACTIONS
3.2.3.1.1 Manufacturing Fluids and Processes
The ,Huids _r l_rocesses used during mum(facturing operations shall m_t
t/uaccCl)lal)l.l' corrode la/tk material, e/llbt'itlle it, or otherwise make il sltSCCl_lihle
to ./)'act/ire or to leakage.
All substances, particularly fltfids and fluid processes to be used in contact with
high-strength tank alloys during manufactttrir_,g operations, storage, or transportation,
including the types of materials and processes described previously, should be evaluatd for
possible incompatibility reactions or effects. This requirement must be observed for
titanium alloys and high-strength alloy steels. Previous successful use of fluids or known
t'ormulatk)rls or processes is one kind of verification data; however, changing conditions of
usage can invalidate such experience and require further evaluation, particularly when the
danger is internal contamination of material, which cannot normally be determined by
nondestructive methods.
The chemical constitutents of all fluid formations or products used on alloys that are
sensitive to corrosion or contamination should be known and kept under control. New
products considered for use should be investigated for new chemical constituents or
markedly different balances of previously used chemicals. The compatibility of new
applications of chemicals or formulations, and that of new fluid products for which the
chemical constituents cannot be determined, should be verified by appropriate test or
investigative procedures before the material is used in contact with high-strength alloys.
Contamination of titanium with chloride-containing products or with hydrogen, oxygen, or
nitrogen during material processing, heat treatment, or welding must be avoided. Lists of
cleauing and processing materials, inspection fluids, testing fluids, and marking and
identifying materials that are considered compatible with titanium (on the Apollo program_
are provided in reference 34. This reference also contains a list of materials fouild to be
incompatible with titanium.
Processes that arc known to contaminate steels with hydrogen must be avoided unless it is
certain that effective embrittlement relief can be provided (normally by baking). The
sources of hydrogen contamination of steel are discussed in reference 36 (pp. 4 through 8)"
the elimination of hydrogen from steel is discussed in reference 38. Atmospheres for heat
treatment of steel must be selected and controlled to prevent hydrogen contamination,
carburization or decarburization, and unwanted scaling.
99
3.2.3.1.2 Testing Fluids
Fluids used in tank and system testing shah not degrade tank materials or react
with them in an unacceptable way.
It is recommended that all fluids to be used in contact with tank materials be maintained
free from undesirable contaminants from any source, including decomposition within the
fluid. The material threshold stress intensity for crack growth in the proof-testing fluidshould be known and taken into consideration in the fracture-mechanics-based safe-life
analysis of the tank (sec. 3.2.4.2).
3.2.3.1.3 Stored Fluids
Fluids stored or contained in a tank shall not decompose, deteriorate, become
contaminated or react with tank materials in any unacceptable way.
The compatibility of tank materials with fluids to be stored must be verified under
conditions of exposure representative of the anticipated service life. The types of test data
that may be required to demonstrate absence of material degradation include exposures ofboth parent metal and welds under stressed and nonstressed conditions. ("Stressed
exposure" requirements are discussed further in sections 3.2.3.2 and 3.2.3.3.) The exposure
parameters that properly should be represented by substantiating data or experience include
exposure temperature, stress, exposure duration, metal heat-treat and surface condition,
fluid composition (including the composition range of important constituents), and(possibly) fluid pressure.
Tank materials should be evaluated for possible deteriorating effects on propellants to be
stored, either by catalytic decomposition or by other types of metal/fluid reactions. This
evaluation is particularly necessary for propellants that tend to be chemically unstable, e.g.,hydrogen peroxide.
3.2.3.1.4 Atmospheric and Environmental Corrosion
Tank materials shall not corrode or otherwise deteriorate below allowable limits
as a result of exposure to atmosphere or other external environments.
Protection from external corrosion, when required, should be provided in accordance with
acceptable guidelines such as those in references 44, 45, 136, and 137, consideration beinggiven to any special environments (such as space environments) or requirements notconsidered in the referenced documents.
100
3.2.3.2 TYPES OF MATERIAL/FLUID REACTIONS
3.2.3.2.1 Stress-Corrosion Cracking
Tank materials shall not experience unacceptable stress-corrosion cracking
resulting from the effects of sustained stress and exposure to testing fluids, stored
fluids, or atmospheric environments.
To prevent stress-corrosion failures, the following guidelines and procedures associated with
tank design, manufacture, and testing should be observed:
(1) Tank alloys must be determined to be free from indications of stress corrosionwhen in contact with fluids to be contained and at stresses up to the material
yield strength, in both parent metal (unwelded) and weld areas.
(2) Threshold stress intensities for flaw growth in the fluid to be contained should be
considered in the tank safe-life analysis (sect. 3.2.4.2).
(3) Fluids selected for tank and system testing must be known to be free from any
tendency to cause stress corrosion when in contact with tank materials.
(4) Materials exposed to atmospheric environments should be resistant to stresscorrosion in the environment, as indicated by the absence of cracking tendencies
up to a sufficiently high threshold stress.
(5) The exposure of material end grain to an aggressive environment should be
avoided, as a general rule, because reduced stress-corrosion resistance in this graindirection is common.
{6) Design details that result in permanent or long-duration high tensile stresses at
exposed metal surfaces should be avoided if possible. Examples are interference
fits and clamped rigid fittings.
(7) Stress raisers and rough surfaces on exposed surfaces should be avoided.
(8) Fabrication procedures should not result in high unrelieved residual stresses.
(9) Coatings or finishes should not be relied on for stress-corrosion protection untiltheir effectiveness has been established.
101
3.2.3.2.2 Galvanic Corrosion
Tank materials shall not deteriorate unacceptably or .fail as a result of galvanic
corrosion.
Galvanically dissimilar metals should not be used in contact or in close proximity in areas
exposed to moisture-laden atmosphere. The immersion of such dissimilar metals in water or
aqueous solutions should be avoided. The permissible metal couples indicated in references44 and 45 should be observed.
When it is necessary or desirable to use metal combinations that have a greater galvanic
potential difference than permitted by the above references, metal corrosion by galvanic
action may be minimized or avoided by observing one or more of the following rules:
(1) Electrically insulate the dissimilar metals from each other, or provide a
sufficiently long path for current flow so that the current is attenuated by electricresistance.
(2) Provide suitable coatings to isolate the metals from the fluid. It is important to
coat both anode (corroded electrode) and cathode because of the possibility that
small defects in the coating will result in unfavorable anode-to-cathode area ratio.
(3) Avoid combinations involving a large cathode area and a small anode area
immersed in the electrolyte. In such cases, the rate of attack on the anode will be
greatly increased.
(4) When practical, provide chemical inhibitors to the fluid.
Nonaqueous liquids to be stored should be evaluated as to electrical conductivity and
development of electrode potentials in contact with the tank metals so that the probability
of galvanic corrosion may be established. Such evaluations should take into account all
contaminants anticipated.
3.2.3.2.3 Hydrogen-Environment Embrittlement
Tank materials used in hydrogen systems shall not ./dil as a result of
hydrogen-environment embrittlement.
Materials that are to be used in contact with pure hydrogen gas should be evaluated for
sensitivity to hydrogen-environment embrittlement. Considerable data on the sensitivity of
various alloys to this failure mode are available in the literature (refs. 65 through 68).
Materials that do not show the effect or have shown very little sensitivity to hydrogen gas
should be selected, particularly for systems in which exposure occurs under pressures greater
102
than oneatmosphere.Whenthereis concernasto tile possibleseverityof the effect, theactualmaterialsandformsto beusedshouldbeevaluatedby obtainingrepresentativetestdata.In suchcases,weldsshouldbeevaluatedaswellasparentmetal.In someapplications,it maybedesirableto usegaseoushydrogenfor pressuretestingof tanks:thispracticewilldemonstratetile reliability of tanksconstructedfromhigh-strengthalloys,whichmayshowsomesensitivityto ahydrogenenvironment.
Titaniumalloysarenot recommendedfor applicationsthatinvolvestructuralloadingwhilethemetalis in contactwith hydrogengasat temperaturesabove I00° F.
3.2.3.2.4 Material Ignition
Prol_elhmt ta_zk materials shall not ignite or otherwise react violeHtly in the
prese_zce o.F the propel�an t.
Titanium and titanium alloys should not be used in contact with oxidizers such as red
fuming nitric acid, liquid oxygen, pressurized gaseous oxygen, mixtures of liquid oxygen
and liquid fluorine, and other strong oxidizers. The possibility of ignition due to energy
pulses from impact, rupture, friction, electricity, heat, or any other source of high localized
energy should be considered and evaluated before titanium or titanium alloys are used in
contact with oxidizers. Copper, lead, zinc, molybdenum, and many alloys that contain these
elements are not compatible with hydrazine and the hydrazine family of propellants. The
compatibility of metals with liquid and gaseous phases associated with propellants,
particularly the highly reactive propellants, must be established before they are used in
contact with such fluids.
The compatibility of any nonmetallic material with a propellant, particularly a strong
oxidizer, must be verified by appropriate test data before the material is used in contact
with such a fluid.
3.2.4 Fracture Control
Tattks shall have a high reliability against brittle fracture in proof test and in
service life.
A complete fracture-mechanics analysis should be performed for each tank to determine
material-toughness requirements, to determine flaw sizes that must be discovered
nondestructively, and to establish proof-testing requirements (refs. 2, 3, and 4). This
analysis should be combined with both an adequately developed program for ensuring
repeatable quality in materials processes and manufacturing operations and a system for
documenting all information pertinent to a tank's structural performance. Such a program
103
will minimize hardware rejection, minimize failures during proof testing, and provide asuitable level of reliability against brittle failures in service.
Catastrophic failures during proof testing should be minimized by the proper correlation of
material toughness in the proof-testing fluid, the material stress at proof pressure, and the
flaw-detection capabilities of the applied nondestructive inspection. The largest crack (or
crack-like flaw) that potentially could escape nondestructive inspection should be
established. The material minimum stress intensity for flaw growth in the proof-testing fluid
(either Ktc or KT n ) should also be determined. This stress intensity value must be adequateto resist the growth, at the proof stress, of the largest crack escaping detection. If it is not,
one or more of the parameters-maximum crack size, material minimum toughness, or
proof stress - should be adjusted to obtain reliability against failure in proof test.
Brittle failures during service should be minimized by proper material selection and
processing (sec. 3.2.4.1), by the performance of a safe-life analysis (sec. 3.2.4.2), and by
implementation of a system of tank documentation, or "pedigree". The documentation
maintained for individual tanks should provide material traceability through all
manufacturing operations, back to the original material procurement and associated
acceptance test results. The results of all inspections performed should be available, togetherwith the details of all repair of rework necessitated. Detailed records of all tank tests or
periods of pressurization, including the fluids contained at such times, should complete thisdocumentation.
3.2.4.1 MATERIAL FRACTURE TOUGHNESS
Tank materials shall have adequate fracture toughness; adequate resistance tocrack growth under sustained loading in service environments, and known flawgrowth characteristics under cyclic loading conditions.
Tank materials should have adequate fracture toughness, under the conditions of use, in
both parent metal and welds. Factors that affect this property (e.g., test temperature, test
direction with respect to material grain, material variation, and production processingeffects) should be taken into account in the selection and evaluation of tank materials. The
toughness of actual production material should be verified by performing tests on welded
and unwelded material that has been processed in the same way as production parts. Quality
control should be provided to ensure adequate toughness in manufactured hardware.
Sustained-load crack growth. To verify material suitability for the intended use and to
allow performance of a complete safe-life analysis (sec. 3.2.4.2), stress intensity values forthe growth of cracks under sustained loading in anticipated fluid environments should be
obtained for both parent metal and welds. The ratios between stress-intensities for crack
growth in the fluids under scrutiny (KTn wdues) and K_c values from comparative tests in
104
dry air (or othersuitablyinert environment)shouldbecomputed.Low valuesof thisratioindicate the possibleincompatibilityof the material-fluidcombinationbecauseof stresscorrosionor someother mechanism.Suchsituationsshouldbe the occasionfor furtherstudyof thematerial-fluidinteraction,both for parentmetalandfor welds.
Cyclic-load crack growth. The material crack-growth characteristics under cyclic loadingin the anticipated fluid environment should be determined, so that crack growth that may
occur during repeated pressurization or other cyclic loads after proof testing and before
critical service loading can be computed as described in section 3.2.4.2.
3.2.4.2 SAFE-LIFE ANALYSIS
A safe-life analysis based on fracture mechanics shall verify tank reliability in
service.
The data utilized for this computation should include the stress intensity for crack growth
in the material in the service fluid and at the service temperature, the cyclic crack-growth
characteristics in the environments anticipated for such loading cycles, the (typical} critical
stress intensity applicable to proof-testing conditions,-and the material stresses associated
with service, cyclic loading, and proof testing. At present, it is considered unconservative touse minimum values for the material critical stress intensity when computing the largest
crack that can exist in a tank after proof test and that can then grow during cyclic loading.
Typical Kit values are more realistic for this computation and for use in determining the
service KTH, an appropriate value for the ratio K,rH/K_ being selected from current
published data for the material under consideration.
The possibility of significant subcritical crack growth during the proof-test cycle should be
investigated, because of the potential effect on the ability of the safe-life analysis toguarantee tank reliability. This requirement applies particularly to low- and
medium-strength materials that have good toughness.
3.2.4.3 ADDITIONAL ELEMENTS IN A FRACTURE-CONTROL PROGRAM
Quality control, qualification testing, and documentation are treated in section 3.2.4.
105
3.3 TANK STRUCTURAL DESIGN
3.3.1 Safety Factor
The safety factor fbr vehicle- and subsystem-tank structure shall be the minimum
required to obtain the desired reliability.
Within the current state of the art, it is not considered advisable to recommend a specific
value of design safety factor. The design safety factor should be based on the reliability
requirements of the specific program. However, certain guidelines are recommended, asfollows:
(I) A base factor of safety should be established for both vehicle and subsystem
tanks. Variations from the base should be justified on the basis of a thorough
knowledge (supported by tests, as required) of the failure mode and the cost,
weight, and reliability effects of these variations in the base factors of safety.
(2) Instability modes of failure (e.g., bulkhead failure under collapsing pressures, or
tank sidewall general instability failure under tank wall compressive loads) should
have a design factor of safety greater than base. For pressure-stabilized
monocoque sidewalls such as those in the Atlas vehicle and for situations where
the primary loading condition is a bending load, postbuckling strength should be
considered in the determination of a safety factor. Stable, tension modes of
failure (e.g., burst pressure failure of tank membrane or bulkheads) may have
design factors of safety less than base.
(3) Welds should have a factor of safety greater than base, because of the
inconsistency in the welding process, variations in the presence and size of
defects, and the high cost of inspection and repair coupled with the small weightpenalty incurred.
(4) Factors of safety should be applied to strain, rather than to stress, in those areas
where the material inevitably is stressed locally into the plastic range or where the
weight penalty to preclude plastic behavior (ref. 122) would be excessive (e.g., the
membrane material immediately adjacent to local "hard spots" areas of greatlyincreased thickness).
(5) For non-ASME-coded tanks, the test programs and quality requirements should
be established to satisfy the Range Safety Criteria.
106
3.3.2 Loads Analysis
The taltk loads pr_,/ile shall iltchtde all iluliridtml desiglt loads or the w_r,s't
comhiJmtiott thereo/i
All axisylnmetric and h)cal design loads, including dynamic loads, should be resolved into
membrane loads to determine the critical load condition. The critical membrane loading
condition should be expressed in terms of a load-temperature-time profile. This profile
should be prepared by plotting all loads and associated temperatures imposed (during
handling, storage, assembly, installation, and service use) versus time. The worst
combination of loads as indicated by the profile then should be used in the design structural
analysis. The load considerations for subsystem tanks internally mounted within a vehicle
should include internal pressure, loads at supports, and related interface plumbing as
amplified by the expected vibration, acceleration, and other flight environments. The load
considerations for integral vehicle tanks should include internal pressure, compressive loads
from weight of tile upper vehicles under flight conditions, side loads due to thrust
misalignment during vehicle direction changes, and local membrane loads imposed by
associated feed, vent. and service lines under flight conditions.
3.3.2.1 TANK SIDEWALL
The loads l_ro/ile Jbr tattk sidewall, s' shall ittc'hlde pressttre, inertial force, axial
l(_od_, dtld hc'HdiJlg tltotttcItts.
Ultimate design tension Ioadings in the longitudinal direction should be determined by
combining ultimate loading due to body axial load and bending moment with ultimate
loading from ullage and head presstire. The ultimate design compressive loading should be
determined by combining ultimate loading due to body axial load and bending with loading
from tank ullage and head presstire; use methods such us those presented in reference 90.
3.3.2.2 END CLOSURE
The loads profile jot tank end closures shall illclude the effect o.f presszlre a_zcl
gL'()tllg try.
The bulkhead geometry should be considered as a means of reducing the hoop loads at tile
tank wall junction. These reduced hoop loads minimize future problems resulting from high
stresses in the juncture. Loading should be determined by methods such as those presented
in reference 97.
107
3.3.2.3 INTERTANK BULKHEAD
The loads profile .for #ttertank bulkheads shall inchtde the effect of pressuredifferentials (burst and collapse) and tempera ture gradients.
The effects of pressure and temperature differentials on common bulkheads should be
determined. The loadings should include maximum temperature gradients produced during
the tank fill operation when both tanks contain cryogenic fluids. An internal-loads analysissuch as the program for a multilayered shell of revolution described in references 98 and 99
should be performed.
When separate membranes are involved, treat the individual bulkheads as end closures (see.3.3.2.2).
3.3.2.4 ATTACHMENT
Attachment loads shall include the radial load, moment, and torque imposed byattachment subsystem.
A loads analysis should be performed to establish the magnitude of loads imposed byattached structure, i.e., line systems, valves, disconnects, and similar attachments. Because
of the complexity of these dynamic analyses, it is recommended that a dynamic-responsecomputer program be used to calculate loads. See reference 138 for a typical analyticalmethod.
3.3.3 Membrane Thickness
The tank membrane thickness shall be the min#num consistent with usage
conditions, reliabi#o' requirements, and fabrication limitations.
For the conventional tank shapes such as spheres and cylinders, use standard formulas ( ref.
100, pp. 299-307) for calculating membrane thickness based on the tank loads (pressure),
shape, and material properties in the expected environments. Apply the factor of safety tothe limit load and establish membrane thickness by using the equations for meridional and
hoop stresses. The greater thickness resulting from use of these equations establishes the
minimum membrane dimension for the intended pressure/environment usage. Themembrane thickness requirements to preclude unacceptable flaw growth (sec. 3.2.4) thenshould be determined. The larger membrane thickness is the dimension to be used fordesign.
108
3.3.4 Sidewall
Vehicle-tattk sidewall structure shall be capable of trattsmittittg the required loads
while containitzg a givett liquid uttder stated c<mditi<ms o.t"accelerati<m, pressure,
arm flight em,irotmtettt.
The recommended procedure for selecting the optimum sidewall configuration is to pert'otto
a trade study along the lines indicated in table VIII. Each item is given a point value
representing an assessment of the relative importance of the item. Each candidate
configuration then is evaluated on how well it rates on each item compared with the
maximum number allowed for that item. The largest total sum for a given candidate
indicates the most desirable configuration. It should be demonstrated that the selected
sidewall design is consistent with the requirements for structural attachment and will
withstand the stress imposed by the bulkhead on the sidewall.
3.3.4.1 SKIN-STRINGER-FRAME
Skin-strittger-t}'ame cottstructio_l ,/or tank sidewalls shall transmit axial loads
without httcklittg.
Skin-stringer-frame construction should be used when large axial loads (e.g., those in the
Saturn IC booster) are applied to the tank sidewall. Trade studies should be conducted to
establish the optimum stringer configuration (blade vs T-stringer). Integrally machined
stringers are preferable because they preclude the need for mechanical attachments that
penetrate the skin and create potential leak paths and stress raisers.
3.3.4.2 WAFFLE
Wa./.lte constructiott .for tank sidewalls shall transmit axial loads and hemtittg
momettts withottt httcklittg.
When axial compressive loads are of moderate magnitude (e.g., those in the Saturn S-IV-B
stage), waffle construction should be employed. In addition, because of its ability to resist
buckling during compressive loadings, waffle construction should be considered when
insulation is required. However. waffle construction is not recommended for pure
membrane loadings as developed in subsystem tanks.
109
Table VIII. - Sample Trade Study for Selection of Optimum
Sidewall Configuration
Tradeoff factor
Technical
Weight
Safety
Reliability
State of the art
Qual./verif. testing
Impact on other systems
Growth potential
Anticipated problemsPerformance
Point
Value
100
30
40
30
40
20
30
20
40
Subtotal 350
Fabrication
Ease of manufacture
Fab. state of the art
Inspection capability
Facilities impact
Hardware availability
45
35
15
30
25
Subtotal 150
Operations
Service equipment impact
Maintainability
Checkout impact
Launch facilities impact
50
30
20
50
Subtotal 150
Cost/schedule
Nonrecurring costs
Recurring costs
Schedule compatibility
5O
150
150
Rating of
configuration
Subtotal 350
Total score 1,000 ! ]
II !II
110
3.3.4.3 MONOCOQUE
Pre,_'sure-,_talfilized mono('oque constrtt(7i(m for tank sidewalls shall transmit axial
loads wiHmul Duckling.
Pressure-stabilized monocoque tanks should be considered when axial loads are less than
bending loads. The membrane can be allowed to buckle partially above limit load if bending
is the predominant loading lref. 122). Consideration should be given to methods of
supporting the structure when the tank is unpressurized.
3.3.5 End Closure
Tank end closure.s" and intertank hulkhea_Ls' shall provide requo'ed tank ./'hdd
calmcity and strttctttral capability withi/t target wei,_,ht and height.
In determining the optimum btdkhead height, consideration should be given to the
following factors:
• Effect of changes in tank height on overall vehicle bending l'llonlents
• The desirability, from a manufacturing and reliability standpoint, of tlsil'lg the
same bulkhead shape throughout a given stage
• Bulkhead deflection under load
• Bulkhead ability to sustain its own weight or other external loads without
collapsing
• Other system space requrements between the bulkhead and skirt (e.g., space
needed for black boxes).
Specific practices for achieving optimum design are set forth in sections 3.3.5.1 through
3.3.5.3 below.
3.3.5.1 FORWARD BULKHEAD
l;(_rward bulk�wad geometry shall be optimized with adjoining stru('mre t()
minimize vehicle height and weight.
The following procedure outlines the recommended method for determining the bulkhead
height that results in the ininimum total weight design:
111
(1)
(2)
(3)
(4)
Studies of w_rious bulkhead shapes should be made, and a curve of bulkhead
weight vs height constructed for each shape. It is recommended that only those
shapes that preclude hoop compressive stresses be considered.
The skirt structure should be optimized for the design running load, and its
weight per inch of length computed.
The tank structure should be similarly treated, and again a weight per inch of
length computed.
The optimum height to achieve minimum total weight of skirt, bulkhead, andtank wall to contain a fixed volume then should be determined by plotting a
curve of total weight vs height (fig. 11).
3.3.5.2 AFT BULKHEAD
Aft bulkheads shall withstand the internal loads resulting from pressure,
acceleration, and the attachment of engines and system components.
The large pressure differential measured from the apex to the equator of an aft bulkhead
dictates the requirement for a compressive stable design in the equatorial region. Waffle-type
structure is recommended in the areas of compressive loads: monocoque construction is
recommended in the noncompressive loads area. A pressure-stabilized aft bulkhead may
carry engine thrust loads if internal support structure to distribute the load is provided; a
semimonocoque structure with an internal thrust barrel as used in the Centaur vehicle is a
good approach.
3.3.5.3 INTERTANK BULKHEAD
An intertank bulkhead shall impose minimum weight on the vehicle tanks.
The intertank bulkhead should be shape optimized to minimize total vehicle weight in the
same manner as the forward bulkhead, as described in section 3.3.5.1.
When extremely lightweight design is necessary and no insulation between fluid tanks is
required, the single-membrane divided bulkhead is recommended. Since the single
membrane has structural capability only under tensile loads, the use must be weighed
against reliability requirements and the undesirability of operational restrictions.
If insulation is required, an insulation bulkhead can be attached adjacent to the structural
bulkhead to form a cavity, as shown in figure 22. The cavity may be filled with insulation or
may be a wtcuum chamber. When vehicle-tankage length is restricted severely, a
112
sandwich-typecommonbulkheadis recommended.This constructionwill reducestagelengthup to one-thirdof tankdiameter,but thebenefitshouldbeweighedagainstincreasedfabricationcomplexityof thebulkheadandthebulkhead-to-sidewalljuncture.
3.3.5.3.1 Sandwich Construction
Sandwich construction used in bulkhead design shall withstaIM tank pressures,
temperature d_[[erentials, and tatzk .fhtM loads.
Sandwich construction (fig. 10) should be optimized to establish the proper core type, size,
and depth and the facing-sheet thicknesses. Reference 102 defines methods that may be
used for this optimization. Consideration should be given to joint design and attachment
provision when honeycomb sandwich construction is selected as the bulkhead material.
3.3.6 Attachment Junctures
3.3.6.1 WELD JOINTS
Weld joints in tank structure shall result in minimttm heat-affected zone, shall
require minitmtm postweM cleaning, at_d shall maintaiJt structural contimdty
uHder all conditions of loading.
The circumferential welds should be removed as far as possible from the bulkhead-sidewall
intersection so that bending stresses at the discontinuity are attenuated. The weld joint
should be designed to operate at a stress and strain level that minimizes the need for repairs,
due consideration being given to the probable ranges of porosity and inclusions that
normally result from welding.
Structural weld joints in tank structure should be butt welded; however, lap-seam welds
with backup rows of spot welds may be used, as was done on the thinwalled Atlas and
Centaur tanks. The preweld joint configuration should be established through consideration
of membrane material and basic thickness. Recommendation of specific configt, rations is
inadvisable because of variations from design to design. It is appropriate, however, to list the
following guidelines that have provided successfltl designs:
(1) Backup rings to prevent dropthrough should be avoided because of resultant
contaminant traps and difficulty of contamination removal.
(2) For subsystem tanks of monocoque design with a membrane thickness of 0.030
in. and less, burndown butt joints are recommended.
113
(3)
(4)
For subsystem tanks of monocoque design with a membrane thickness greater
than 0.030 in., a groove and filler-wire joint is recommended.
For thin-walled tanks up to 18 in. diameter, the weld-joint transition (taper)
length on both sides of the weld centerline should cover an arc of at least 30 °
originating at ti_e tank centerline (fig. 25).
Section A-A
bsystem tamk
Figure 25. - Sketch of geometry of a weld joint for a subsystem
tank.
(5)
1(7)
(7)
Where possible, weld lines should be located so that the joint transition (taper)
does not extend into the hemispherical section.
Oll subsystem tanks where the exterior skin surface necd not be smooth, the
majority of the transition material should be outside the membrane median line
so that weld joint "sink in" is offset.
Weld lands should be joined to the basic tank membrane by liberal transition
sections and fillet radii (fig. 13(b)L
t14
3.3.6.2 BULKHEAD/SIDEWALL JUNCTURE
Attachment /uttctures s/tall provide reliable, leakprc)qf paths ./br the loads .from
the comtecti_tg major eomponeHts (skirts, tatzk sidewalls, altd bulkhead) apM shall
mi_timize disco_ttimtit.v stresses.
Welding the tankage enclosure is the best way to avoid leakage. Y-ring designs shown in
figure 15 arc recommended for avoiding leakage and providing at] efficient structural load
path. There shonht be sufficient material in the Y-ring itself to preclude excessive hooptension stresses in the meridional weld.
Bulkhead n]embrane thickness and tank stringers should be tapered as they approach the
circumferential welds, so that bending moments caused by eccentricities arc reduced.
Eccentricities in the skirt-to-Y-ring outer leg also should be minimized.
Rigorous analysis of major junctures must be verified by testing under all critical design
loading conditions.
Because of the difficulty of repair, the undesirability of tank reentry and cleaning, the
necessity of postrepair retesting, and the schedule impact caused by replacement of parts
where further weld repair is not feasible, it is highly recommended that weld design be very
conservative.
3.3.6.3 BOSSES AND SUPPORT PROVISIONS
Ta_tk bosses a_ld support provisions shall impose minimum discontimdties in the
laHk met_lbraHes.
Structure for attachments should be integral and SlnOothly blended into the basic tank
membrane so that stress and strain concentration are minimized. Two configurations are
shown in figure 26, configuration B being the recommended design. Fillet radii should be
large, and transition pads liberal and preferably tapered. An alternate approach is the
welded-in pad or ring; however, the design choice of a welded-in ring versus an integrally
milled ring for membrane bulkheads should be evaluated carefully for each particular
application, the reliability and freedom-from-repair characteristics of the integral design
being balanced against the general acceptability and much lower cost of the welded-in
design.
115
._ j_Hard"spot
Configuration A
I
Configuration B
Figure 26. - Two configurations of structure for attachments to
a tank.
3.3.7 Openings and Access Doors
Tank openings and access doors shall limit leakage to acceptable levels and shall
result in minimum additional local stresses and strains on the tank.
The bolt attach ring and adjoining portion of the door or tank opening should be designed
to minimize joint rotation and to match as nearly as possible the deflection of the tank
structure if 11o opening existed. The effects of eccentricities in flange load should be
minimized by tapered or stepped transition sections.
Because each new tank design usually has its own unique port requirements, it is inadvisable
to recommend specific configurations. Good design practices, however, require
consideration of certain guidelines applicable to all configurations, as follows:
(1) Ports should be integral with the tank membranes.
(2) Threaded bosses should be employed as a tank opening where possible tominimize discontinuities in tank membranes.
(3) On threaded bosses in thin-walled tanks up to 18 in. diameter, the transition
(taper) length should be at least as long as an arc within a 20 ° angle originating at
the tank centerline (fig. 27).
(4) Threaded bosses should have external wrenching pads to facilitate installation of
mating fittings.
116
_/Boss
/Transition taperTank membrane
200
min.
Figure 27. - Recommended geometry for transition from boss to
tank membrane.
(51 Ports should be as large as possible to facilitate inspection of the tank interior.
(6) Closures for access openings should be made of material with quality equivalent
to the tank material.
17) Closures for access openings should minimize bending in the adjacent tank
rile mbrane.
(8) Tank port provisions should be located and designed to simplify forging dies.
(9) Threaded inserts should have bolt locking provisions, and the bolting flanges
should be capable of repair in the event of thread damage.
117
3.4 TANK COMPONENTS
3.4.1 Propellant Slosh and Vortex Suppression Devices
Baffle configuratioH and location shall prevent adverse liquid motion.
General recommendations for baffle configuration are not possible since various approaches
appear to work satisfactorily for a particular condition. A comprehensive study of slosh
suppression and slosh loads is presented in references 103 and 139, respectively. Two basic
guidelines are as follows:
(1) In a vehicle with a requirement for multiple engine starts, the baffles should be
installed in the tank at a position slightly below the anticipated surface level of
the liquid at the time that suppression is required (e.g., at engine cutoff).
(2) Vanes or baffles should be located around a tank fluid exit such that fluid swirling
is disrupted without causing cavitation of the downstream pump.
3.4.2 Propellant Positioning Devices
The propellant positioning device shah provide the required liquM control under
flight em,irottmeHts and conditions, and its weight and volume sttall be minimal.
The propellant positioning device should have the required structural capability and rigidity
to withstand the dynamic loads imposed by propellant behavior and by llight enviromnents.
The design should have minimum impact on the associated tank design in terms of
installation and removal needs and support requirements within the tank.
The propellant positioning device should be capable of being cleaned, flushed, and dried
when fully assembled in the tank.
The screen mesh size should be as large as possible to facilitate manufacturing and cleaning
and to minimize the possibility of hole clogging by particulate contamination.
Partial-fluid-control devices should be sized and located so that the propellant quantity
retained will sustain engine firing sufficiently long to ensure propellant settling within thetank.
118
3.4.3 Propellant Expulsion Devices
The expulsiott device s/tall continuously maintain the .fluid to be expelled at the
taJt k outlet under all conditions when Olt t/'low iS required.
Selection of an expulsion device should be based upon consideration of the tank size, weight
constraints, and expected usage. The tank mountings and the expulsion-device attachment
points should be as closely coincident as possible to minimize vibration amplification into
the expulsion device.
Expulsion devices that contain the fluid should be capable of evacuation preparatory to
fluid servicing, or the tank must be oriented with a vent port oriented upward.
If a bladder-type device is to be used in vertically mounted cylindrical tanks, a slightly
undersize cylindrical section in the bladder to reduce frictional drag between the bladder
and tank wall should be considered.
3.4.3.1 EXPULSION EFFICIENCY
The expuMon device shall have maximum expulsion efficiency.
For maximum expulsion efficiency, the liquid to be expelled should be contained within the
device. If a flexible bladder is used, the central feedout tube should be as small as possible to
minimize fluid quantity that the bladder cannot expel. It is recommended that the tube be
designed as a long flexible spring, so that under "g" forces it can follow the motions of the
fluid. The tube flow area also should be sized to introduce negligible flow resistance.
If a metal device is used, it should be designed with nesting convolutes to minimize residual
fluid that may remain lodged in each convolute. The moving closeout head of the bellows
should be designed to nest into tile tank contour to improve volumetric efficiency of the
tank system.
For piston-type devices, tile piston seals should be designed for reliability against particulate
contamination, wear, and fluid damage, so that negligible leakage will occur during service
life.
The pressure required to actuate the expulsion device fully should be as low as possible so
that consistent fluid flow throughout the expulsion cycle is ensured.
119
3.4.3.2 MATERIAL
The expulsion-device material shall satisfy all compatibility, leakage, andtemperature requirements and shall be of minimum weight.
No single material is suitable for all expulsion-device applications. To select the optimummaterial, the designer must identify all operating conditions in the order of their
significance. The material that can withstand or fulfill the most usage requirements usually
is the proper choice. It can be seen, however, that establishing the comparative importance
of the various usage conditions depends heavily on tile judgement of the designer.
Selection of material for an expulsion device must take into account the interrelation of the
functional and environmental requirements and the material characteristics. For extended
missions or usage periods (e.g., ten years) compatibility and resistance to permeation shouldhave more significance in material selection than, for example, ease of fabrication. For
extremely cold or warm environments, the selection should be based heavily on materialresistance to degradation at temperature extremes.
If permeation of liquid vapors and pressurant gas across the expulsion device is unacceptable
to upstream system components or engine operation, a positive barrier device such as ametal bellows or diaphragm or bladder with a metallic barrier should be used.
The designer should ensure that the material selected is compatible with all fluids that will
be used throughout the life of the expulsion device. Particular emphasis should be placed on
the assessment of fluids that may be employed while the device is subjected to operatingstresses.
3.4.3.3 DESIGN MARGIN
The expulsion device shall have an adequate capability beyond the intendedoFerational functions and service environment conditions.
Establishment of a combined operational and environmental program for demonstration of
the design margin of an expulsion device requires caution and judgement on the part of the
designer. The imposition of combined requirements that are unrealistically severe and
damaging to the expulsion device should be avoided. A liberal margin should be imposed
only on those requirements that have a high degree of uncertainty. For example, the fill and
expulsion cycles can be predicted with reasonable accuracy and, therefore, cyclerequirements beyond anticipated usage should be minimized. Mission temperature and
vibration specifications, on the other hand, usually have a degree of uncertainty during thedevelopment phase, and a requirement for demonstrating a more liberal margin in theseareas is justified.
120
The possibleeffectsof combinedoperationaland environmentalconditionssuchas fullworking pressureat elevatedtemperatureshould be evaluatedso that prematureandprobablyuninformativefailure is precluded.Thedesignershouldalsoattemptto evaluatethe added affect of gravity on expulsion device perl'ormance when themargin-demonstrationtestsare conducted.This precautionis particularlysignificantforhorizontallymountedcylindricaltanksthatemploybladder-typeexpulsiondevices.
3.4.4 Tank Insulation
TaHk i_lsulation shall withstand the tank strains resulting from temperature,
pressure, and vehicle body loads.
The selection of an insulation system for cryogenic tank application should consider the
magnitude of the applied strains. When strain levels are high, spray foam or bolted insulationis recommended. To avoid the possibility of insulation damage from ice formation and from
handling, internal insulation should be used when the improved material properties at
cryogenic temperatures are not involved in the attempt to achieve minimum weight.Methods for repairing localized insulation damage should be included in the insulation
design considerations.
3.5 TANK DESIGN ANALYSIS
A design analysis shall veriJ), the structural acceptability of the tank design.
Both structural and dynamic analyses should be performed concurrently with the design.Consideration should be given to tensile and compressive stresses arising from pressure loads,
thermal loads, and static and dynamic loads, particular attention being given to major
junctures and local attachments and openings.
3.5.1 Strength Analysis
Analysis by accepted analytical techniques shall verify tank structural integrity
jbr all critical flight and ground conditions.
Margins of safety should be computed and specified for all structural elements. The
analytical methods used should be conservative to the extent of the uncertainties in the
analytical methods or the manufacturing processes. All assumptions should be clearly stated,
121
andextremecautionshouldbeexercisedin theformulationof theanalytical model usedto simulatetheactualstructure,especiallyif analysisisperformedby computer.Areasthatrequire verification by testing should be so indicated.The analysisshould includedeterminationof theeffectsof combined loading as well as cyclic and sustained loading.
3.5.1.1 TENSION-LOADED STRUCTURE
Tensile stresses a,M deilections in the shell structure shall not exceed allowable
values for yieM arid rupture under the combined loads fi)r all critical designconditions.
Deflection calculations should be based on nominal material thickness. For yield andultimate stress calculations, minimum thickness should be used.
Analysis of welds should be empirical and based on test results for the particular weld land
configuration being used. Test procedures must be determined carefully so that both the
test-specimen fabrication and loadings represent a proper simulation of real hardware.
Because of the difficulty and expense in performing biaxial tests, it is highly recommended
that a realistic set of weld specifications be firmly established early in a program so that
design allowables for welds and weld repairs are consistent with fabrication, inspection, andrepair procedures on the actual hardware.
3.5.1.2 COMPRESSION-LOADED STRUCTURE
Compressive stresses on the shell structure shall not exceed allowable values foryield and buckling under the combined loads for all critical design conditions.
The analytical calculations should be based on nominal dimensions. In general, instability
computations should reflect conservative values and should be followed by full-scale testingthat simulates the critical biaxial load and thermal conditions.
3.5.1.3 MAJOR JUNCTURES
Discontinuity stresses from critical combined bending and axial loads at tank
ma/or /unctures shall not exceed material allowables for rupture.
The analysis should include all elements that make up a juncture and should be based on
normal material tolerances. Analysis should confirm that circumferential welds are located
in regions where joint discontinuity stresses are minimum; if not, sufficient material shouldbe provided to reduce the discontinuity effect.
122
3.5,1.4 LOCAL ATTACHMENTS AND OPENINGS
The design stress at membrane local attachments and openings shall _zot exceed
the allowable yield or ultimate stress.
A finite-element computer program similar to that shown in reference 125 is recommended
for the analysis of the reinforced openings and pads. This computer program will handle a
shell structure of arbitrary geometry and loading. A qualification test of typical openings
and attachments should be performed to verify the design.
3.5.2 Structural Dynamics
The tan k structure shall withstand all transient and steady-state dynamic loads or
the worst combination of d.vnamic loads and critical static loads.
Detailed dynamic analysis of the particular stage and the vehicle should be performed to
ensure that the tank's design is adequate for all imposed transient and steady-state dynamic
loads. The dynamic loads imposed on the tank as determined from the individual dynamic
analysis should be integrated into the vehicle structural analysis. The axial, shear, and
bending distribution resulting from transient dynamic loading conditions should be
compared with equivalent static loading conditions and should be included in the vehicle
load-time-temperature-history profile, The transient dynamic stresses should be combined
with any static or steady-state vibratory stresses when applicable.
Recommendations for specific methods of analysis for all dynamic conditions are beyond
the scope of this monograph; however, dynamic analysis techniques that may be used are
discussed in references 140 through 143.
For dynamic analysis of clustered structures, matrix techniques or continuous-mechanics
methods (refs. 131 through 134 and 144 through 147)may be used.
3.5.2.1 BENDING FREQUENCY
The tank body [_ending frequency shall be within the limits imposed by the
vehicle flight control system or by predicted transient dynamic loads.
The vehicle-tank stiffness, including El, G J, and AE distributions, where
123
E = modulus of elasticity
I = moment of inertia
G = modulus of rigidity
J = torsion constant
A = area of cross section
should be defined and used in the dynamic-model analysis of the vehicle.
Vehicle-tank stiffness should be consistent with tile minimum stiffness required to ensure
stable aeroelastic behavior of the vehicle, to ensure structural adequacy under transient
dynamic loads, and to limit the body-bending frequencies to within the capabilities of theguidance and control systems (refs. 148 and 149).
3.5.2.2 EXTERNAL DYNAMIC ENVIRONMENT
Tanks shall withstand the maximum transient longitudinal and transverse flight
loads and the shipping and handling loads.
The dynamic analysis for these conditions should include the dynamic characteristics of the
portion of vehicle remaining at any point in flight and the characteristics of the test standwhen applicable.
The determination of the shear and bending dynamic loads should include tile vehicle
dynamic characteristics (natural frequency in bending) and the harmonic content of the
forcing function. Analysis of dynamic interaction among the guidance and control system,the TVC system, and the vehicle also should be included.
Procedures for shipping and handling of vehicles should require suitable packaging and
harness supports to limit the transient dynamic loads imposed during handling and shipping
to within the load capability of the vehicle tank as designed for flight. The dynamic
characteristics of any suspension system and any shock or vibration mitigation systems
included in the handling equipment or shipping container should be included in the dynamic
analysis of the tank for transportation and handling environments.
124
3.6 TANK FABRICATION
The tank and component fabricatioH processes shall be the most reliahle, the least
time consuming, and the most cost effective )'or the particular tank and program
needs.
An engineering study of fabrication processes should he accomplished to select the
fabrication processes that afford the best compromise between fabrication schedule and
costs without reducing reliability below specified levels. The engineerin_ study should
include detailed tradeoff evaluations of fabrication and welding processes; past experience
with and reliability of the various processes; schedule effect of the processing; and
fabrication, tooling, and facility costs versus the tank configuration.
Fabrication processes should be selected carefully to avoid harmful effect on the material
and end product. This selection requires a detailed analysis of the effect that fabrication
processes will have on the material and the completed tanks. If not available, information on
process effects should be developed in a material- and process-evaluation program.
3.7 TESTING AND INSPECTION
Testing shall be adequate to evaluate the basic' taHk design, and t/re inspection
processes shall be capable of detecting the unacceptable defects in tank materials
and in the fabricated tank.
it is not possible to make across-the-board testing recommendations, since each tank
program has its own design and usage conditions that dictate unique testing. Destructive
testing of full-scale tanks, particularly of large vehicle tanks, often is prohibitive from a cost
standpoint. However, destructive testing of properly designed subscale tanks should be
conducted when necessary for evaluating the full-scale tank. Subscale test tanks must be
designed to duplicate the following parameters of the full-scale tank:
• The wall thickness (burst stress on the subscale membrane equal to the expected
burst stress of the full-scale case)
• Production materials
• Production methods and processes
• Inspection methods
125
The tank proof test shouldbedesignedon the basis of fracture-mechanics theory (see.
3.2.4). Test pressure, test temperature, external axial and bending loads, and pressurization
rates should be in accordance with specific program requirements.
Inspection processes should be employed throughout the tank program beginning with
material proctlrement and continuing through fabrication, process control, and final
acceptance. Each phase can use different inspection techniques with different acceptance or
rejection standards. For this reason, an overall master plan for the use and management of
the quality-control program should be established prior to the start of fabrication. The
scope of the master plan should be established on the basis of the required reliability level,
the type and orientation of defects encountered, and the process sensitivity required.
126
APPENDIX A
Conversion of U.S. Customary Units to S! Units
Physical quantity
Density
Energy
Force
Fracture toughness
Length
Mass
Pressure
Temperature
Tensile stress
Volume
Yield strength
U.S. customaryunit
lbm/in. 3
ft-lbf
lbf
ksi-in.'/2
SI unit
kg/m 3
N-m
N
(N/m2).m '/_
Conversion
factor a
2.768X 104
1.356
4.448
1.099× 106
in.
mil
Ibm
atm
psi (lbf/in. 2)
ksi (1000 psi)
oF
ksi
ft 3
gal
ksi
cm
/am
kg
N/m 2
N/m 2
N/m 2
K
N/m 2
m3
m3
N/m 2
2.54
25.4
0.4536
l.O13XlO s
6895
6.895X 10 6
5K = - (°F + 459.67)
9
6.895X 106
28.32× 10- 3
3.785X10 -3
6.895X 106
aMultiply value given in U.S. customary unit by conversion factor to obtain equivalent valuein SI unit. For a complete listing of conversion factors, see Mechtly, E. A.: The InternationalSystem of Units. Physical Constants and Conversion Factors. Second Revision, NASA SP-7012, 1973.
12 7
128
Term or Symbol
A
ACS
APS
a
acr
ai
allowable load
(or stress)
alpha
be ta
C
CM
C
Charpy impactstrength
combined stresses
coupon
APPENDIX B
GLOSSARY
Definition
area of cross section
attitude control system
auxiliary propulsion system
crack dimension of primary interest (usually, maximum crack depth)
critical crack dimension for unstable propagation at a given stress
initial crack dimension
load (or stress) that, if exceeded, produces tank failure. Failure may be
defined as buckling, yield, or ultimate, whichever condition prevents
the tank from performing its function.
designation for the microstructure of titanium and its alloys when the
structure is hexagonal close-packed
designation for the microstructure of titanium and its alloys when the
structure is body-centered cubic
material constant in evaluating crack growth
command module (Apollo spacecraft)
one-half the length of a part-through crack
impact strength measured in a test in which a notched bar (of specified
dimensions) is struck by a swinging pendulum; the energy absorbed inthe fracture is measured. A striking velocity of 17.5 ft/sec is employed;
test values are given in ft-lbf
stresses resulting from simultaneous action of all loads to which a
structure is subject
a piece of metal representative of a batch, mill run, or lot, from which a
metallurgical test specimen is prepared
129
Term or Symbol Definition
creep
cryogenic
d
delta
design burst
pressure
design safety factor
design ultimate
load
design yield load
E
ELI
FH
Fbru
Fbry
Fcy
Fsu
Ftu
Fry
G
slow but continuous deformation of a material under constant load or
prolonged stress
fluids or conditions at low temperatures, usually at or below -150 ° C
weld land width
a change in a quantity (e.g., an increase in volume)
maximum limit pressure multiplied by the ultimate factor of safety
an appropriate arbitrary multiplier greater than 1 applied in design to
account for design contingencies such as slight variations in material
properties, fabrication quality, load magnitude, and load distributions
within the tank structure
limit load multiplied by the ultimate design safety factor
limit load multiplied by tile yield design safety factor
modulus of elasticity
extra low interstitial
full hard temper
design ultimate bearing strength
design bearing yield strength
design compressive yield strength
design ultimate shear strength
design ultimate tensile strength
design tensile yield strength
modulus of rigidity
acceleration due to gravity
130
Term or Symbol Definition
HAZ
1
J
K
Kc
Kit
KT H
L/D
limit load
limit pressure
margin of safety (MS)
lnenrb rane
n
NDI
OAMS
heat-affected zone
moment of inertia
torsion constant
crack-tip stress-intensity factor
critical stress-intensity factor
stress intensity for initiation of unstable crack growth under conditions
of maxinmm constraint as in thick sections (plane-strain fracture
toughness)
threshold stress intensity: highest stress intensity for which there is no
crack growth under sustained load in a given enviromnent
length-to-diameter ratio
maximunl expected load that will be experienced by the tank structure
under the specified conditions of operation, with allowance for
statistical variation
maximum pressure that will be experienced by the tank structure under
specified conditions of operation. Maximum limit pressure is the
maximum vent valve pressure plus hydrostatic head (if applicable);
minimmn limit pressure is taken as tile minimum operating pressure of
the tank under the specified conditions of operation, plus hydrostatic
head (if applicable).
fraction by which the allowable load or stress exceeds tire applied load
or stress, 1MS=-- - 1
R
tank skin or shell
material constant in evaluating crack growth
nondestructive inspection
orbital attitude and maneuvering system
131
Term or Symbol
operating pressure
P
pH
proof pressure
Q
R
£y
RCS
SM
SPS
STA
T3,T4,T6,T8 ,T73,
T76,T87
TIG
TVC
t
ullage
ultimate load (or pressure)
Definition
nominal ullage pressure to which the tanks are subjected under
steady-state conditions in service operations
pressure
negative logarithm of hydrogen ion concentration, a measure of acidity
or alkalinity
maximum limit pressure (q.v.) multiplied by the proof-test safety
factor. Proof pressure is the reference from which the pressure levels for
acceptance testing are established.
flaw shape and plasticity parameter, Q = _2 _ 0.212 (o/Oys) 2
(1) ratio of minimum to maximum stress intensity during cyclic
loading
(2) ratio of the design load (or stress) to the allowable load (or stress)
plastic-zone radius of surface crack
reaction control system
service module (Apollo spacecraft)
secondary propulsion system
solution treated and aged
designations for heat-treating and tempering processes for aluminum
alloys
tungsten-inert-gas (welding method)
thrust vector control
material thickness
amount that a container lacks of being full
load (or pressure) at which catastrophic failure (general collapse or
rupture) of the tank structure occurs
132
Term or Symbol Definition
ultimate stress
W
W
XFH
yield load
yield stress
A
12
p
a
Oe ff
o"x , Oy, o z
Oys
stress at which the material fractures or becomes structurally unstable
weight
width
extra full hard temper
load that must be applied to the tank structure to cause a permanent
deformation of a specified amount
stress at which the material exhibits a permanent deformation of
0.0020 inch per inch (0.2 percent)
angle designating the location of K along crack front
incremental change in a quantity
Poisson's ratio
material density
applied stress; in a cracked specimen, stress remote from the crack
effective normal stress
normal stresses acting on three mutually perpendicular planes of zeroshear stress
material yield strength
complete elliptic integral of the second kind,
,_a,2 _ {c2 _ a2} I= 1- sin2 0 dOo kc 2 1
133
Material t
(designation in monograph)
Identification
Metals
A286
AM-350
AM-355
PH 13-8Mo
PH 14-8Mo
PHI5-7Mo
15-5PH
17-4PH
17-7PH
21Cr-6Ni-9Mn
300 series
1100
2014
2024
2219
3003
D6Ac
300M
4130
4140
4335V
4340
5052
5083
5086
5456
heat-treatable, high-strength austenitic steel
semi-austenitic or martensitic precipitation
hardening stainless steels
and transformation
austenitic stainless steels
wrought aluminum (99% AI min.)
wrought aluminum alloys with copper as principal alloying element
wrought aluminum alloy with manganese as principal alloying element
high-strength martensite-hardening low-alloy steels
wrought aluminum alloys with magnesium as the principal alloying
element
1Additional information on metallic materials herein can be found in tile 1972 SAE Handbook, SAE, Two Pennsylvania
Plaza, New York, N.Y.; in MIL-HDBK-5B, Metallic Materials and Elements for Aerospace Vehicle Structures, Dept. ofDefense, Washington, D.C., Sept. 1971 ; and in Metals Handbook (8 th ed.), Vol. 1 : Properties and Selection of Metals, Am.
Society for Metals (Metals Park, Ohio), 1961.
134
Material(designation in monograph)
6O61
7O75
7079
|tS 188
ItS 25(l.605)
ttY 140
9Ni-4Co-0.20C
9Ni-4Co-0.25C
lnconel 718
low-alloy steel
tnaraging steel
Waspaloy
Nonmetals
butyl rubber
Cerrobend
Dacron
Kapton
Mylar
Nonlex
Identification
wrought aluminunl alloy with magnesium and silicon as the principal
alloying elements
wrought aluminutn alloys with zinc as the principal alloying element
cobalt-base Ifigh-temperature superalloys
martensite-hardening special category steels
trade name of International Nickel Co. for precipitation-hardening
nickel-chrommm-iron alloy
steel with low carbon content
martensile- and age-hardening nickel-iron alloy
designalion of Pratt & Whitney Division of [Inited Aircraft Corp. for a
precipitation-hardening nickel-base superalloy
synthetic rubber produced by copolymerization of isobutene with a
small proportion of isoprene or butadiene
trade name of Cerro Sales Corp. for the eutectic alh,y of bismuth, lead,
tin, and cadmium;m.p. 158 ° F.
trade name of E.I. duPont Co. for a polyester fiber made from
polyethylene lerephthalale
trade name of E.I. duPont Co. for a polyimide film (I to 5 mils thick)
trade name of E.I. duPont Co. for a polyester film made from
polyethylene terephflmlate
trade nalne of E.I. duPont Co. for a high-temperature aromatic
polyamide
135
Material(designationin monograph)
polyurethane
Te fl on (T F E )
Teflon (FEP)
TFE/FEP laminale
Fluids
A-50
CRES
FLOX
Freon
helmm
hydrazine
H2 O2
IRFNA
LIt:
LOX
MMH
MON 10
nitrogen
Identification
any of various flrermoplastic polymers that contain-Nit('()()-linkages;
produced as fibers, coatings, flexible and rigid foanls, ela_,tomers, and
resins
trade nanre of E.I. duPont (;o. for polytetrafluoroethylene
trade name of E.I. duPont Co. for a polymer of fluorinated
ethylene-propylene
bladde, made from numerous spray coatings of Teflon (TFE) on a
mandrel with heal cure between each coat followed by numerous spray
coatings of Tefhm (FEPL also wilh heat cure between spray, coals.
Process was developed by Dilectrix Corp., Farmingdale, N.Y.
50/50 blend of hydrazine and UI)MII per MIL-P-27402
corrosion-resistant steel
mixture of liquid fluorine and liquid oxygen
trade name of E.I. duPont Co. for a family of fluorinated hydrocarbons
pressurant helium (He) per MIL-P-27407
N 2 H4, propellant grade per MIL-P-26536
hydrogen peroxide
inhibited red fuming nitric acid, propellant grade per MIL-P-7254
liquid hydrogen (tt:), propellant grade per MI L-P-27201
liquid oxygen (02), propellant grade per MIL-P-25508
mononrethylhydrazine, propellant grade per MIL-P-27404
mixed oxides of nitrogen (905_ N2 04/10% NO)
gaseous nitrogen per MIL-P-27401
136
Material(designation in monograph)
N_lt4
N()
N2 04
RFNA
RP- I
Lq)Mll
Abbrevia tions
ABMA
AFFI)I,
A[:MI.
AIAA
AS('E
ASMt{
ASIM
CPI A
I)MI('
lit
IPC
N1SIC
TAC()M
Identification
see hydra/the
nitric oxide
nilrogen telroxide {oxidizer), propellant grade per MII,-P-2_53 _)
red fuming nilric acid
kerosene-base high-energy hydrocarbon fuel, propellant grade per
MIL-P-2557(_
unsymmeirical dimethylhydra.,'ine, prcq'_ellant grade per .MlIA"-25e,04
Identification
Army Ballistics Missile Agency
Air Force Flight Dynamics Laboratory
Air Force Materials Laboratory
American Institute of Aeronautics & Astronautics
American Society c,f Civil Engineers
American Society _1 Mechanical Engineers
American Society for Testing and Materials
Chemical Propulsi_m Inl'ormati_m Agency
l)efen se Met als Ill formation Center
Illinois Institute of Technology
Lockheed Pr_qmlsion Co.
Marshall Space Flighl Cenler
(Army) Tank-At, lomotive ('ommand
137
AbbreviationsIdentification
WADD
WPAFB
Wright Air Development Division
Wrighl-Palterson Air Force Base
138
REFERENCES
1. Anon.: Aerospace Tanks. Vols. I and II. Rep. Contract NAS 7-388, liT Research Institute, 1969.
2. Liu, A. F.: Fracture Control Methods for Space Vehicles. Vol. I - Fracture Control Design Methods.
NASA CR-134596, Space Div., Rockwell International Corp. (to be published).
3. Ehret, R. M.: Fracture Control Methods for Space Vehicles. Vol. 1I- Assessment of Fracture
Mechanics Technology for Space Shuttle Applications. NASA CR-134597, Space Div., Rockwell
International Corp. (to be published).
4. Anon.: Fracture Control of Metallic Pressure Vessels. Space Vehicle Design Criteria Monograph,
NASA SP-8040, May 1970.
*5. Boddy, J. A.; Mitchell, J. C.; Martindale, C. W.; and Moss, W. L.: Users' Manual Vehicle and
Structural Design Synthesis Program. SD 67-1204-2, Space Div., North American Rockwell Corp.,
Jan. 1968.
6. Shanley, F. R.: Weight-Strength Analysis of Aircraft Structures. Second ed., Dover Publications, Inc.,
1960.
7. Bert, C. W.; and Hyler, W. S.: Design Considerations in Selecting Materials for Large Solid-Propellant
Rocket Motor Cases. DMIC Rep. 180, Battelle Memorial Institute (Columbus, Ohio), 1962, pp.
46-55.
8. Anon.: Metallic Materials and Elements for Aerospace Vehicle Structures. Military Standardization
Handbook, MIL-HDBK-5B, Dept. of Defense, Sept. 1971.
9. Anon.: Plastics for Aerospace Vehicles. Military Standardization Handbook, MIL-HDBK-17A, Dept.
of Defense, Sept. 1970.
10. Hood, D. W.; Eichenberger, T. W.; and Lovell, D. T.: An Investigation of the Generation and
Utilization of Engineering Data on Weldments. Tech. Rep. AFML-TR-68-268, The Boeing Co., Oct.
1969.
I 1. Anon.: Low-Temperature Creep Characteristics of Ti-5AI-2.5Sn and Ti-6AI-4V Alloys. DMIC Tech.
Note, Battelle Memorial Institute, June 8, 1964.
12. Kiefer. T. F.; and Schwartzberg, R. F.: Investigation of Low-Temperature Creep in Two Titanium
Alloys. Final Report, Contract NAS9-5842, Martin Co., June 1967.
*Dossier for design criteria monograph "Liquid Rocket Metal Tanks and Tank Components". Unpublished. Collected
source material available for inspection at NASA Lewis Research Center, Cleveland, Ohio.
139
13.Anon.:MechanicalPropertiesat VariousTemperatures.AlcoaResearchLaboratories,AhuninumCompanyofAmerica,Aug.1962.
14.Rawe,R.E.;andCorn,D.L.: A Comparisonot"theExperimentalBiaxialStrengthof StructuralAlloyswithTheoreticalPredictions.J.Materials,vol.4,no.1,Mar.1969,pp.3-18.
15.Baht,G. K.: A Study of tile Behavior of Small Pressure Vessels under Biaxial Stress Conditions and ill
the Presence of Surface Cracks. Final Report Contract NOnr-3764 (oo) (X), The Mellon Institute,
June 25, 1963.
16. Anon.: Solid Rockel Motor Metal Cases. Space Vehicle I)esign Criteria Monograph, NASA SP-8025,
Apr. 1970.
17. Anon.: Aluminum, Volume III Fabrication and Finishing. American Society for Metals (Metals
Park, Ohio), 1967.
18. Gerds, A. F.: Strohecker, D. E.: Byrer, T. G.;and Bouljer, F. W.: Deformalion Processing of Titanium
and its Alloys. NASA TM-X-53438, Apr. 1%6.
19. Strohecker, D. E.: Gerds, A. F.; ttenning, It. J.: and Bouljer, F. W.: Deformation Processing of
Stainless Steel. NASA TM-X-53569, Oct. 1966.
20. Strohecker D. E.; Gerds, A. F.; and Bouljer, F. W. Deformation Processing of
Precipitation-Hardenable Stainless Steels. NASA TM-X-53431, Oct. 1965.
21. Strohecker, D. E.; Byrer, T. G.; Gerds, A. F.; Gehrke, J. tt.; and Bouljer, F. W.: Deformation
Processing of Nickel-Base and Cobalt-Base Alloys. NASA TM-X-53439, Apr. 1966.
22. Taylor, L., ed.: Metals ttandbook. Volume 4 Forming. Eighth ed., American Society for Metals,1969.
23. Nolan, N. C.; Gadberry, H. M.: Loser, J. B.' and Sneegas, E. C.: Itigh-Velocity Metal Working. NASA
Tectmology Utilization Survey, NASA SP-5062, 1967.
24. Anon.: Welding i|andbook. Sixth ed. (published in five sectionst, American Welding Society (Uniicd
Engineering Center, 345 East 471h St., New York, NY 10017), 1968.
25. Linnert, G. E.: Welding Metallurgy, Carbon and Alloy Sheets. Vols. 1 and 2. Third ed., American
Welding Society (New York), 1967.
26. Thielsch. H.: Defects and Failures in Pressure Vessels and Piping. Reinhold Publishing ('_rp. (New
York), 1965.
27. Anon.: Heat Treatment of Ah,minun_ Alloys. Military Specification MIL-H-6088, Dept. of Defense,Mar. 24, 1965.
28. Anon.' tteat Treatment of Steels. Military Specification MIL-II-_)875, l)ept, of l)efense. Mar. 3t,1967.
140
*29. Anon.:Ileal Treatmentof TilanimnAlloys.SpecificationMA0011-014,SpaceandInformationSystemsDiv.,NorthAmericanAviation,Inc.,July19(_5.
30.Taylor,L., ed.: Melalst|andbook.Vol. 2 Ileal Treating,CleaningandFinishing.Eighthed.,AmericanSocietyforMetals,1964.
31.Williams,D. N.: Ilydrogenin TitaniumandTitaniumAlloys.TMLRep.10(1,BattelleMemorialha_slitute,May195S.
32.Wood,P,.A.;Williams,D.N.;andOgden,tt. R.: Investigationof Delayed-CrackingPhenomenoninHydrogenatedUnalloyedTitanium.DMICMemorandum124,BattelleMemoriallnslitute,Aug.1961.
33. Ferguson,J.hi.: The Oxidation and Contamination of Titanium and its Alloys. DMIC Memorandum
23g, Battelle Memorial Institute, July 1<)68.
Anon.: Compatibility of Materials with Titanium. Specification MB0110-008, Space Div., North
American Rockwell Corp., Apr. t970.
Anon.: Prevention and Elimination of liydrogen Embrittlemenl in Steel. Process Specification
MA0111-016, Space and Information Systems Div., North American Aviation, Inc., May 1965.
36. Elsea, A. R.: and Fletcher, E. E.: The Problenr of Itydrogen in Steel. DMIC Memorandum 180,
Battelle Memorial Institute, Oct. 1963.
37. Elsea, A. R.; and Fletcher, E. E.: llydrogen-lnduced, Delayed, Brittle Failure of High-Strength Steels.
DMIC Rep. 106, Battelle Memorial hlslitute, Jan. 1964.
38. Fletcher, E. E.; and Elsea, A. R.: tlydrogen Movement in Steel Entry, Diffusion, and Elimination.
DMIC Rep. 219 Battelle Memorial Institute, June 1965.
*39. Brownfield, C. D.: The Stress Corrosion of Titanium in Nitrogen Tetroxide, Methyl Alcohol, and
Other Fluids. Rep. SD 67-213A. Space Div., North American Rockwell Corp., June 1968.
*40. Leverton, C.; and Goltbrath, J.: Failure Analysis Addendum of Apollo Launch Escape Motor Cases
ED-204 and ED-20_ during Hydrotest at Excelco Developments, Inc.. Rep. LPC 801-44521A,
Lockheed Propulsion Co., Div. Lockheed Aircraft Corp., Apr. 1968.
41. Anon.: Liquid Propellant Manual. CPIA, .lohns Hopkins University (Silver Springs, Md.) (various
dates).
42. Boyd, W. K.: Berry, W. E,; and White, E. L.: Compatibility of Materials with Rocket Propellants and
Oxidizers. DMIC Memorandum 201, Battelle Memorial Institute, Jan. 1965.
*Dossier for design criteria monograph "Liquid Rocket Metal Tanks and Tank Components". Unpublished. Collected
source material available for inspection at NASA Lewis Research Center, Cleveland, Ohio.
*34.
*35.
141
43.
44.
45.
*46.
*47.
48.
49.
50.
51.
52.
53.
Anon.:DissimilarMetals.MilitaryStandardMIL-STD-889,Dept.ofDefenseSept.1969.
Anon.:ProtectiveFinishingSystemsfor Rockets,GuidedMissiles,SupportEquipment,andRelatedMaterials.MilitaryStandardMIL-STD-186,Dept.ofDefense,Mar.1964.
Anon.:Finishingof MetalsandWoodSurfaces.MilitaryStandardMIL-STD-171,Dept.of Defense,Feb.1964.
Anon.:FinishSpecification,ApolloSpacecraft.Dwg.No.V14-00024,SpaceandInt'ormationSystemsDiv.,NorthAmericanAviation,Inc.,Apr.1965.
Anon.:FinishSpecification,SaturnS-If Stage.SpecificationMA0621-00t,SpaceDiv.,NorthAmericanRockwellCorp.,May1968.
Anon.: The Stress-CorrosionCrackingand Hydrogen-StressCrackingof Metals.MilitaryStandardizationHandbook,MIL-HDBK-724,Dept.ofDefense,Oct.1969.
Logan,H.L.:TheStressCorrosionofMetals.JohnWiley& Sons,Inc.,Sept.1966.
JacksonJ.D.;andBoyd,W.K.:CorrosionofTitanium.DMICMemorandum218,BattelleMemorialInstitute,Sept.1966,
Anon.:PropellantGradeNitrogenTetroxide.SpecificationNASAMSC-PPD-2,Aug.t968,AmendedDec.1968.
O'Brien,R.E.:PrematureEnvironmentalStressCrackingofTitaniuminMethanol,Freon,andOtherSolutions.DMICMemorandum228(AcceleratedCrackPropagationof Titaniumby Methanol,HalogenatedHydrocarbons,andOtherSolutions),BattelleMemorialInstitute,Mar.1967.
Tolberg,W.E.;Rewick,R.T.; andDolder,M.P.:ChemicalandMelallurgicalAnalysesof 6AI-4VTitaniumTestSpecimensExposedto ttydrazine(N2H4)Liquid Propellant. Rep. 951581-11
(prepared for Jet Propulsion Lab.), Stanford Research Institute (Menlo Park, Calif.), Apr. 1971.
54. Setterlund, R. B.: Stress Corrosion Cracking of High-Strength Alloys. Rep. 2684 (Final),
Aerojet-General Corp., Sept. 1963.
*55.
56.
Groenveld, T. P. Elsea, A. R. and Hall, A. M.: The 9Ni-4Co Steels. DMIC Memorandum 220, Battelle
Memorial Institute, Oct. 1966.
Slunder, C. J.: Stress-Corrosion Cracking of tligh-Strength Stainless Steels in Atmospheric
Environments. DMIC Rep. 158, Battelle Memorial Institute, Sept. 1961.
*Dossier for design criteria monograph "Liquid Rocket Metal Tanks and Tank Components". Unpublished. Collectedsource material available for inspection at NASA Lewis Research Center, (' eveland, Ohio.
142
57. Hoenie,A. F.;andRoach,D.B.:NewDevelopmentsinHighStrengthStainlessSteels.DMICRep.223, Battelle Memorial Institute, Jan. 1966.
58. Wagner, H. J.; Burus, R. S.; Carroll, T. E.; and Simon, R. C.: Nickel Base Alloys - Alloy 718. DMIC
Processes and Properties Handbook, Battelle Memorial Institute, Feb. 1968, pp. 1-4.
59. Copson, H. R.; and Cheng, C. F.: Some Case Histories of Stress Corrosion Cracking of Austenitic
Stainless Steels Associated with Chlorides. Paper presented at the Twelfth Annual Conference,
National Association of Corrosion Engineers (St. Louis, Mo.), Mar. 1957. (May also be available from
the International Nickel Co., Inc., Bayonne, N.J.).
60. Logan, H. L.; and McBee, M. J.: Stress-Corrosion Cracking of Cold-Reduced Austenitic Stainless
Steels. Materials Research and Standards, vol. 7, no. 4, Apr. 1967, pp. 137-145.
61. Boyd, W. K.: Stress Corrosion Cracking of Titanium and Its Alloys. DMIC Menlorandum 27_4
(Titanium for the Chemical Engineer), Battelle Memorial Institute, Apr. 1968, pp. 35-44.
62. Anon.: Stress-Corrosion Cracking of Titanium. ASTM-STP-397, American Society for Testing and
Materials, 1966.
63. Tiffany, C. F.; and Masters, J. N.: Investigation of the Flaw Growth Characteristics of 6AI-4V
Titanium Used in Apollo Spacecraft Pressure Vessels. NASA CR-65586, 1967.
64. Bannerton, N. P.; DeRespiris, D. L.; and Vogt, J. W.: Carbonation Cell Materials Compatibility.
NASA CR-1481, Dec. 1969.
65. Campbell, J. E.: Effects of Hydrogen Gas on Metals at Ambient Temperature. DM1C Rep. S-31,
Battelle Memorial Institute, Apr. 1970.
66. Walter, R. J.; and Chandler, W. T.: Effects of High Pressure Hydrogen on Metals. ASM Rep. D8-14.2,
American Society for Metals, Oct. 1968.
67. Walter, R. J.; Hages, H. G.; and Chandler, W. T.: Mechanical Properties of lnconel 718, Waspaloy,
A-286 and Ti-5A1-2.5Sn ELI in Pure Gaseous Hydrogen. Rep. R-8187, Rocketdyne Div., North
American Rockwell Corp., Apr. 1970.
68. Lorenz, P. M.: Effect of Pressurized Hydrogen upon Inconel 718 and 2219 Aluminum. Final Report,
Aerospace Group, The Boeing Company, 1969.
69. McPherson, W. B.; and Cataldo, C. E.: Recent Experience in High Pressure Gaseous Hydrogen
Equipment Operated at Room Temperature. Tech. Rep. D8-14, American Society for Metals, Oct.
1968.
70. Williams, D. N.; and Maykith, D. J.: Reaction of Titanium with Gaseous Hydrogen at Ambient
Temperatures. DMIC Tech. Note, Battelle Memorial Institute, Feb. 1966.
143
71. Koehl, B. G.; Hodge, W.; Williams, D. N.; and Bartlett, E. S.: Continuation of the Investigation of the
Reaction of Titanium with Hydrogen. Rep. Contract NAS 9-6565, Battelle Memorial Institute, May1967.
72. Koehl, B. G.; Williams, D. N.; and Bartlett, E. S.: Investigation of the Reaction of Titanium with
Hydrogen. NASA CR-92389, Mar. 1969.
73. Pittinato, G. F.: The Effects of a Hydrogen Environment on the Mechanical Properties of ForgedTi-6A1-4V ELI Weldments. NASA CR-102660, Nov. 1969.
74. Rittenhouse, J. B.; et al.: Corrosion and Ignition of Titanium Alloys in Fuming Nitric Acid.
WADC-TR-56-414, Wright Air Development Center (WPAFB, OH), Feb. 1957.
75. Jackson, J. D.; et al.: A Study of the Titanium-Liquid Oxygen Pyrophoric Reaction.WADD-TR-60-258, Wright Air Development Division (WPAFB, OH), Mar. 1960.
76. Jackson, J. D.; Boyd, W. K.; and Miller, P. D.: Reactivity of Metals with Liquid and Gaseous Oxygen.DMIC Memorandum 163, Battelle Memorial Institute, Jan. 1963.
77. White, E. L.; and Ward, J. J.: Ignition of Metals in Oxygen. DMIC Rep. 224, Battelle MemorialInstitute, Feb. 1966.
78. Anon.: Testing of Materials other than Sealants for Compatibility with Liquid Oxygen.AMBA-PD-M-44, Army Ballistic Missile Agency Purchase Description, July 1958.
79. Anon.: Testing Compatibility of Materials for Liquid Oxygen Systems. NASA-SPEC-106, NASAMSFC, Feb. 1964.
80. Rawe, R. E.: Fracture Mechanics and Safe Life Design. DAC 59591, Douglas Aircraft Co., McDonnellDouglas Corp., Aug. 1970.
81. Irwin, G. R.: Crack Extension Force for a Part-through Crack in a Plate. J. Appl. Mech., Trans.ASME, Series E, vol. 84, 1962, pp. 651-654.
82. Brown, W. F., Jr.; and Srawley, J. E.: Fracture Toughness Testing Methods. ASTM STP-381,
American Society for Testing and Materials, 1964.
83. Brown, W. F., Jr.; and Srawley, J. E.: Plain Strain Crack Toughness Testing of High Strength MetallicMaterials. ASTM STP-410, American Society for Testing and Materials, 1966.
84. Anon.: Fatigue Crack Propagation. ASTM STP-415, American Society for Testing and Materials,1967.
85. Wessel, E. T.; et al.: Engineering Methods for the Design and Selection of Materials against Fracture.
Final Report, Contract DA-30-O69-AMC-602(T) (AD 801005), U.S. Army Tank-Automotive Center,June 1966.
144
86. Campbell,J. E.: Plain-Strain Fracture-Toughness Data for Selected Metals and Alloys. DM1C Report
S-28, Battelle Memorial Institute, June 1969.
87. Anon.: Third Annual Workshop in Fracture Mechanics. Universal Technology Corp., 1966.
88. Wilhelm, D. P.: Fracture Mechanics Guidelines for Aircraft Structural Applications.
AFFDL-TR-69-111 (AD 702528), Northrop Corp., Feb. 1970.
89. Smith, H. R.; Piper S. H.; and Downey, F. K.: A Study of Stress-Corrosion Cracking by Wedge Force
Loading. Engineering Fracture Mechanics, vol. 1, no. 1, June 1968, pp. 123-128.
90. Paris, P. C.; and Erdogen, F.: A Critical Analysis of Crack Propagation Laws. J. Basic Eng., Trans.
ASME, Series D, vol. 85, 1963, pp. 528-534.
91. Forman, R. G.; Kearney, V. E.; and Engle, R. M.: Numerical Analysis of Crack Propagation in
Cyclic-Loaded Structures. J. Basic Eng., Trans. ASME, Series D, vol. 89, 1967, pp. 459-464.
92. Wilson, W. K.; Clark, W. G., Jr.; and Wessel, E. T.: Fracture Mechanics Technology for Combined
Loading and Low-to-lntermediate Strength Metals. AD 682754, Westinghouse Research Labs.
(Pittsburgh, Pa.), 1968.
*93. Collipriest, J. E., Jr.: and Kizer, D. E.: Experimental Verification of Proof-Test Logic. Paper
presented at Western Metal and Tool Conference, American Society for Metals (Los Angeles, Calif.),
Mar. 9-12, 1969.
*94. Collipriest, J. E., Jr.: Part-through-Crack Fracture Mechanics Testing. Rep. SD 71-319, Space Div.,North American Rockwell Corp., June 1971.
*95. Anon.: Apollo Command and Service Module Pressure Vessels, General Specification For.
MC999-O022A, Space and Information Systems Div., North American Aviation, Inc., Oct. 1962.
96. Anon.: Stress Analysis of Saturn S-II Stage Airframe. SID 65-1676, Space and Information Systems
Div., North American Aviation, Inc., Jan. 1966.
97. Timoshenko, S.; and Woinowsky-Kreiger, S.: Theory of Plates and Shells. Second ed., McGraw-Hill
Book Co., 1959.
*98. Capelli, A. P.: General Equilibrium Equations for Three-Layered Shells of Revolution. Re p. STR-123,
Space and Information Systems Div., North American Aviation, Inc., Mar. 1965.
*99. Capelli, A. P.: Axisymmetric Deformations for Three-Layered Shells of Revolution. Rep. STR-124,
Space and Information Systems Div., North American Aviation, Inc., June 1965.
*Dossier for design criteria monograph "Liquid Rocket Metal Tanks and Tank Components". Unpublished. Collectedsource material available for inspection at NASA Lewis Research Center, Cleveland, Ohio.
145
100.
*101.
102.
103.
104.
105.
106.
107.
108.
109.
110.
111.
112.
113.
Roark, Raymond J.: Formulas for Stress and Strain. Fourth ed., McGraw-Hill Book Co., 1965, pp.
293-309.
McHugh, A. H.: User's Manual Pressurized Structure Optimization Program RAX Computer
Program "FRAP". Space Div., North American Rockwell Corp., Jan. 1971.
Anon.: Honeycomb Sandwich Design. Brochure E, Hexcel Products, Inc. (Berkeley, Calif.), 1959.
Anon.: Slosh Suppression. Space Vehicle Design Criteria Monograph, NASA SP-8031, May 1969.
DeBrock, S. C.; et al.: A Survey of Current Development in Surface Tension Devices for Propellant
Acquisition. Sixth A1AA Propulsion Joint Specialist Conference (San Diego, Calif.), June 15-19,
1970.
Orton, G. F.: Simulated Flight Vibration Testing of a Surface Tension Propellant Expulsion Screen.
Paper presented at CPIA Meeting, (Las Vegas, Nev.) Sept. 1970.
Lee, J. C.: Counter Permeation Phenomenon in Bladdered Expulsion Tanks. Joint AIAA/Aerospace
Corp. Symposiutn on Low Gravity Propellant Orientation and Expulsion (Los Angeles, Calif.), May1968.
Porter, R. N.; and Stanford, H. B.: Propellant Expulsion Bladder Evaluation. Space Programs
Sumnaary 37-66,Vol. 1. Jet Propulsion Lab., Calif. Inst. Tech. (Pasadena, Calif.), July 1, 1966.
Anon.: Codispersion Propellant Expulsion Bladder Evaluation. Space Programs Summary 37-66,Vol.
1. Jet Propulsion Lab., Calif. Inst. Tech., Sept. 1 to Oct. 31, 1970.
Anderson, F. K.; and Thompson, L. M.: Correlation of Referee Fluid and Propellant in Vibration
Testing. Rep. 8514-928005, Bell Aerospace Co., Nov. 1970.
Cuddihy, E. F.: Solvent-Stress-Cracking and Fatigue Properties of Liquid-Propellant Expulsion Teflon
Bladders. Tech. Rep. 32-1535, Jet Propulsion Lab., Calif. Inst. Tech., Aug. 1971.
Krusos, J. N.; Higgins, R. G.; and Dukes, W. H.: Design Criteria and Quality Control Studies for
Teflon Expulsion Bladders. Rep. 8460-833012, Bell Aerospace Co., Mar. 1967.
Anon.: Space Program Summary 37-36. Jet Propulsion Lab., Calif. Inst. Tech., Oct. 1 to Nov. 30,
1965.
Thompson, L. M.: Welded Metal Bellows, A Reliable Positive Expulsion Device for Liquid
Propellants. AIAA Paper 64-264, First AIAA Annua! Meeting (Los Angeles, Calif.), July 2, 1964.
*Dossier for design criteria monograph "Liquid Rocket Metal Tanks and Tank Components". Unpublished. Collectedsource material available for inspection a_ NASA Lewis Research Center, Cleveland, Ohio.
146
"121.
*125.
114.Koval,L. R.:A Zero-GSloshProblem.J.SpacecraftRockets,vol.5, no. 7, July 1968, pp. 865-868.
115. Leising, C. J.;and Stocky, J. F.: Structural and Mass Transport Phenomena Associated with the Use
of Teflon Expulsion Bladders on the Mariner Mars 1971 Spacecraft. Paper presented at I lth Liquid
Propulsion Symposiunl (Miami Beach, Fla.), Sept. 1960.
116. Lange, R. L.: Evaluation of bletallic Positive Expulsion Bellows Operating Parameters wilh Cryogenic
Propellants. Rep. Contract NAS 3-13327, Bell Aerospace Co., July, 1971.
117. Timoshenko, S.: Strength of Materials. Third ed., McGraw-Hill Book Co., 1955.
118. Peery, D. J.: Aircraft Structures. McGraw-Hill Book Co., 1950.
I I 9. Anon.: Structure Manual. Pub. 2546-N, Space Div., North American Rockwell Corp., Oct. 1969.
120. Anon.: Test Evaluation Report V7-22-1. SD 69-330, Space Div., North American Rockwell Corp.,
July 1969.
Baker, E. It.: Design Procedure for Integrally Stiffened Waffle Cylinders under Axial Compression.
Rep. STR-77, North American Aviation, Inc., May 1961.
122. Peery, D. J.: Post Buckling Strength of a Pressurized Cylinder. GD/A Rep. DDG64-024A, General
Dynamics/Convair, Ocl. 1964.
I|etenyi, M.: Beams on Elastic Foundation. The Universily of Michigan Press, 1946.
Treppo, V. S.; and Mulcahy, E. J.: Flex I and Flex !1 Computer Program Analysis Report.
STR-SS-69-3, North American Rockwell Corp., May 5, 1969.
Modeer, J. R.; and Kovalevsky, L.: Analysis of a Reinforced Circular or Elliptical Cutout. Rep.
STR-173, Space and Information Systems Div., North American Aviation, Inc., Mar. 1967.
126. Bijlaard, P. P.: Stresses in Spherical Vessels from Radial Loads Acting on a Pipe. Bulletin 49, Welding
Research Council, Apr. 1959.
127. Bijlaard, P. P.: Stresses from Local Loadings in Cylindrical Pressure Vessels. Trans. ASME, vol. 77,
1955, pp. 805-816.
128. Bijlaard, P. P.: Stresses from Radial Load in Cylindrical Pressure Vessels. Welding J., Res. Suppl.,
1955, pp. 608-617.
129. Bijlaard, P. P.: Computation of the Stresses from Local Loads in Spherical Pressure Vessels or
Pressure Vessel Heads. Bulletin 34, Welding Research Council, Mar. 1957.
*Dossier for design criteria monograph "Liquid Rocket Metal Tanks and Tank Components". Unpublished. Collected
source material available for inspection at NASA Lewis Research Center, Cleveland, Ohio.
147
130.Clincan,B. E.; Gates,R. M.; andAndrews,J. S.: DynamicLoadsduringBoostedFlight.ASD-TDR-63-302(AD405020),TheBoeingCompany,1963.
131.Loewy,R.G.;andJoglekar,M.M.:MatrixHolzerAnalysisforFully-CoupledVibrationsofClusteredLaunch-VehicleConfigurationsincludingApplicationsto theTitanIII-CandUncoupledSaturnICases.NASACR-592,1966.
132.Duty,J.P.:A Studyto DeterminetheEffectsofClusteringofSolidEnginesonStructuralVibrationLevels.NASACR-68433,1965.
133.Keith,J. S.;et al.: Methodsin StructuralDynanticsfor ThinShellClusteredLaunchVehicles.AFFDL-TDR-64-105,Ling-Temco-Vought,1965.
134.Hurty,W. C.: Dynamic Analysis of Structural Systems using Component Modes. J. Aerospace
Sciences, vol. 3, no. 4, Apr. 1965, pp. 678-685.
135. Anon.: Pressurization Systems Design Guide. Vol. 1. Aerojet General Corp., 1965.
136. Anon.: Finishes and Coatings: General Specification for Protection of Aerospace Weapons Systems,
Structure and Parts. Military Specification MIL-F-7179, Dept. of Defense, Sept. 1965.
137. Anon.: Surface Treatment and Metallic Coatings for Metal Surfaces of Weapon Systems. MilitarySpecification M1L-S-5002, Dept. of Defense, Nov. 1966.
*138. Anon.: 3D Modal Beam Program. Space Div., North American Rockwell Corp., no date.
139. Anon.: Propellant Slosh Loads. Space Vehicle Design Criteria Monograph, NASA SP-8009, Aug.1968.
140.Mixson, J. S.; and Herr, R. W.: The Linear Elastic Dynamic Analysis of Shells of Revolution by the
Matrix Displacement Method. AFFDL-TR-66-80, Proc. Matrix and Structure Mechanics Conference
(WPAFB,OH., Oct. 26-28, 1965), 1966.
141.Klein, S.; and Sylvester, R. J.: The Linear Elastic Dynamic Analysis of Shells of Revolution by the
Matrix Displacement Method. AFFDL-TR-66-80, Proc. Matrix and Structure Mechanics Conference
(WPAFB,OH., Oct. 26-28, 1965), 1966.
142. Przemieniecki, J. S.: Theory of Matrix Structure Analysis. McGraw-Hill Book Co., 1968.
143. Argyrus, J. H.: Energy Theorems and Structural Analysis. Butterworth & Co. (London), 1960.
144. Lianis, G.: Matrix Analysis of Vibration ot" Clustered Boosters. Rep. AE61-0858, GeneralDynamics/Convair, Sept. 1961.
*Dossier for design criteria monograph "Liquid Rocket Metal Tanks and Tank Components". Unpublished. Collectedsource material available for inspection at NASA Lewis Research Center, Cleveland, Ohio.
148
145.Milner,J.L.: Three-Dimensional Multiple Beam Analysis of a Saturn I Vehicle. NASA TM X-53098,
July 1964.
146. Kiefling, L.: Multiple Beam Vibration Analysis of Saturn 1 and IB Vehicles, NASA TM X-53072, June
1964.
147. Meissner, C. J.; and Levy, R. S.: Flexibility Method of Coupling Redundant Complex Structures. J.
Structural Division, ASCE, vol. 89, no. ST6, Dec. 1963.
148. Lukens, D. R.; Schmitt, A. F.; and Broucek, G. T.: Approximate Transfer Functions for
Flexible-Booster-aud-Autopilot Analysis. WADD TR-61-93, Wright Air Development Div. (WPAFB,
OH), Apr. 1961.
149. Greensite, A. L.: Analysis of Liquid Propellant Mode Stability of a Multi-Tank Booster Vehicle, J.
Aeron. Sci., vol. 29, no. 2, Feb. 1962, pp. 130-139.
149
150
NASA SPACE VEHICLE DESIGN CRITERIAMONOGRAPHS ISSUED TO DATE
ENVIRONMENT
SP-8005
SP-8010
SP-8011
SP-8013
SP-8017
SP-8020
SP-8021
SP-8023
SP-8037
SP-8038
SP-8049
SP-8067
SP-8069
SP-8084
SP-8085
SP-8091
SP-8092
Solar Electromagnetic Radiation, Revised May 1971
Models of Mars Atmosphere (1967), May 1968
Models of Venus Atmosphere (1972), Revised September 1972
Meteoroid Environment Model--1969 (Near Earth to Lunar Surface),
March 1969
Magnetic Fields Earth and Extraterrestrial, March 1969
Mars Surface Models (1968), May 1969
Models of Earth's Atmosphere (90 to 2500 kin), Revised March 1973
Lunar Surface Models, May 1969
Assessment and Control of Spacecraft Magnetic Fields, September 1970
Meteoroid Environment Model 1970 (Interplanetary and Planetary),
October 1970
The Earth's ionosphere, March 1971
Earth Albedo and Emitted Radiation, July 1971
The Planet Jupiter (1970), December 1971
Surface Attnospheric Extremes (Launch and Transportation Areas),
May 1972
Tile Planet Mercury (1971), March 1972
The Planet Saturn (1970), June 1972
Assessment and Control of Spacecraft Electromagnetic Interference,
June 1972
151
SP-8103
SP-8105
STRUCTURES
SP-8001
SP-8002
SP-8003
SP-8004
SP-8006
SP-8007
SP-8008
SP-8009
SP-8012
SP-8014
SP-8019
SP-8022
SP-8029
SP-8030
SP-8031
SP-8032
SP-8035
SP-8040
SP-8042
SP-8043
The Planets Uranus, Neptune, and Pluto (1971), November 1972
Spacecraft Thermal Control, May 1973
Buffeting During Atmospheric Ascent, Revised November 1970
Flight-Loads Measurements During Launch and Exit, December 1964
Flutter, Buzz, and Divergence, July 1964
Panel Flutter, Revised June 1972
Local Steady Aerodynamic Loads During Launch and Exit, May 1965
Buckling of Thin-Walled Circular Cylinders, Revised August 1968
Prelaunch Ground Wind Loads, November 1965
Propellant Slosh Loads, August 1968
Natural Vibration Modal Analysis, September 1968
Entry Thermal Protection, August 1968
Buckling of Thin-Walled Truncated Cones, September 1968
Staging Loads, February 1969
Aerodynamic and Rocket-Exhaust Heating During Launch and Ascent
May 1969
Transient Loads From Thrust Excitation, February 1969
Slosh Suppression, May 1969
Buckling of Thin-Walled Doubly Curved Shells, August 1969
Wind Loads During Ascent, June 1970
Fracture Control of Metallic Pressure Vessels, May 1970
Meteoroid Damage Assessment, May 1970
Design-Development Testing, May 1970
152
SP-8044
SP-8045
SP-8046
SP-8050
SP-8053
SP-8054
SP-8055
SP-8056
SP-8057
SP-8060
SP-8061
SP-8062
SP-8063
SP-8066
SP-8068
SP-8072
SP-8077
SP-8079
SP-8082
SP-8083
SP-8095
SP-8099
QualificationTesting, May 1970
Acceptance Testing, April 1970
Landing Impact Attenuation for Non-Surface-Planing Landers, April1970
Structural Vibration Prediction, June 1970
Nuclear and Space Radiation Effects on Materials, June 1970
Space Radiation Protection, June 1970
Prevention of Coupled Structure-Propulsion Instability (Pogo), October
1970
Flight Separation Mechanisms, October 1970
Structural Design Criteria Applicable to a Space Shuttle, Revised March1972
Compartment Venting, November 1970
Interaction with Umbilicals and Launch Stand, August 1970
Entry Gasdynamic Heating, January 1971
Lubrication, Friction, and Wear, June 1971
Deployable Aerodynamic Deceleration Systems, June 1971
Buckling Strength of Structural Plates, June 1971
Acoustic Loads Generated by the Propulsion System, June 1971
Transportation and Handling Loads, September 1971
Structural Interaction with Control Systems, November 1971
Stress-Corrosion Cracking in Metals, August 1971
Discontinuity Stresses in Metallic Pressure Vessels, November 1971
Preliminary Criteria for the Fracture Control of Space Shuttle
Structures, June 1971
Combining Ascent Loads, May 1972
153
SP-8104
GUIDANCEANDCONTROL
SP-8015
SP-8016
SP-8018
SP-8024
SP-8026
SP-8027
SP-8028
SP-8033
SP-8034
SP-8036
SP-8047
SP-8058
SP-8059
SP-8065
SP-8070
SP-8071
SP-8074
SP-8078
SP-8086
StructuralInteractionWithTransportationandHandlingSystems,January1973
GuidanceandNavigationforEntryVehicles,November1968
Effectsof StructuralFlexibilityonSpacecraftControlSystems,April1969
SpacecraftMagneticTorques,March1969
SpacecraftGravitationalTorques,May1969
SpacecraftStarTrackers,July1970
SpacecraftRadiationTorques,October1969
EntryVehicleControl,November1969
SpacecraftEarthHorizonSensors,December1969
SpacecraftMassExpulsionTorques,December1969
Effectsof StructuralFlexibilityonLaunchVehicleControlSystems,February1970
SpacecraftSunSensors,June1970
SpacecraftAerodynamicTorques,January1971
SpacecraftAttitudeControlDuringThrustingManeuvers,February1971
TubularSpacecraftBooms(Extendible,ReelStored),February1971
SpaceborneDigitalComputerSystems,March1971
PassiveGravity-GradientLibrationDampers,February1971
SpacecraftSolarCellArrays,May1971
SpaceborneElectronicImagingSystems,June1971
SpaceVehicleDisplaysDesignCriteria,March1972
154
SP-8096
SP-8098
SP-8102
Space Vehicle Gyroscope Sensor Applications, October 1972
Effects of Structural Flexibility on Entry Vehicle Control Systems,
June 1972
Space Vehicle Accelerometer Applications, December 1972
CHEMICAL PROPULSION
SP-8087
SP-8109
SP-8052
SP-8110
SP-8081
SP-8048
SP-8101
SP-8100
SP-8094
SP-8097
SP-8090
SP-8080
SP-8064
SP-8075
SP-8076
SP-8073
SP-8039
Liquid Rocket Engine Fluid-Cooled Combustion Chambers, April 1972
Liquid Rocket Engine Centrifugal Flow Turbopumps, December 1973
Liquid Rocket Engine Turbopump Inducers, May 1971
Liquid Rocket Engine Turbines, January 1974
Liquid Propellant Gas Generators, March 1972
Liquid Rocket Engine Turbopump Bearings, March 1971
Liquid Rocket Engine Turbopump Shafts and Couplings, September
1972
Liquid Rocket Engine Turbopump Gears, March 1974
Liquid Rocket Valve Components, August 1973
Liquid Rocket Valve Assemblies, November 1973
Liquid Rocket Actuators and Operators, May 1973
Liquid Rocket Pressure Regulators, Relief Valves, Check Valves, Burst
Disks, and Explosive Valves, March 1973
Solid Propellant Selection and Characterization, June 1971
Solid Propellant Processing Factors in Rocket Motor Design, October
1971
Solid Propellant Grain Design and Internal Ballistics, March 1972
Solid Propellant Grain Structural Integrity Analysis, June 1973
Solid Rocket Motor Performance Analysis and Prediction, May 1971
155
SP-8051
SP-8025
SP-8041
Solid Rocket Motor Igniters, March 1971
Solid Rocket Motor Metal Cases, April 1970
Captive-Fired Testing of Solid Rocket Motors, March 1971
] 56 NASA-Langley,1974
4
NATIONAL AERONAUTICS AND SPACE ADMINISTRATION
WASHINGTON, D,C, 20546
OFFICIAL BUSINESS
...... T.... ,3oo SPECIALo FOURTH-CLASS
BOOK
RATE
PC'_TMASTER :If Undeliverable (Section 15_
Postal .M:Hl,l_ll) D() Not Return